agora inbox for pgsql-hackers@postgresql.org  

help / color / mirror / Atom feed

[PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
659+ messages / 2 participants
[nested] [flat]

* [PATCH v5 2/7] Row pattern recognition patch (parse/analysis).
@ 2023-09-02 06:32 Tatsuo Ishii <ishii@postgresql.org>
  0 siblings, 0 replies; 659+ messages in thread

From: Tatsuo Ishii @ 2023-09-02 06:32 UTC (permalink / raw)

---
 src/backend/parser/parse_agg.c    |   7 +
 src/backend/parser/parse_clause.c | 292 +++++++++++++++++++++++++++++-
 src/backend/parser/parse_expr.c   |   4 +
 src/backend/parser/parse_func.c   |   3 +
 4 files changed, 305 insertions(+), 1 deletion(-)

diff --git a/src/backend/parser/parse_agg.c b/src/backend/parser/parse_agg.c
index 85cd47b7ae..aa7a1cee80 100644
--- a/src/backend/parser/parse_agg.c
+++ b/src/backend/parser/parse_agg.c
@@ -564,6 +564,10 @@ check_agglevels_and_constraints(ParseState *pstate, Node *expr)
 			errkind = true;
 			break;
 
+		case EXPR_KIND_RPR_DEFINE:
+			errkind = true;
+			break;
+
 			/*
 			 * There is intentionally no default: case here, so that the
 			 * compiler will warn if we add a new ParseExprKind without
@@ -953,6 +957,9 @@ transformWindowFuncCall(ParseState *pstate, WindowFunc *wfunc,
 		case EXPR_KIND_CYCLE_MARK:
 			errkind = true;
 			break;
+		case EXPR_KIND_RPR_DEFINE:
+			errkind = true;
+			break;
 
 			/*
 			 * There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_clause.c b/src/backend/parser/parse_clause.c
index 334b9b42bd..60020a7025 100644
--- a/src/backend/parser/parse_clause.c
+++ b/src/backend/parser/parse_clause.c
@@ -100,7 +100,10 @@ static WindowClause *findWindowClause(List *wclist, const char *name);
 static Node *transformFrameOffset(ParseState *pstate, int frameOptions,
 								  Oid rangeopfamily, Oid rangeopcintype, Oid *inRangeFunc,
 								  Node *clause);
-
+static void transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static List *transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist);
+static void transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
+static List *transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef);
 
 /*
  * transformFromClause -
@@ -2950,6 +2953,10 @@ transformWindowDefinitions(ParseState *pstate,
 											 rangeopfamily, rangeopcintype,
 											 &wc->endInRangeFunc,
 											 windef->endOffset);
+
+		/* Process Row Pattern Recognition related clauses */
+		transformRPR(pstate, wc, windef, targetlist);
+
 		wc->runCondition = NIL;
 		wc->winref = winref;
 
@@ -3815,3 +3822,286 @@ transformFrameOffset(ParseState *pstate, int frameOptions,
 
 	return node;
 }
+
+/*
+ * transformRPR
+ *		Process Row Pattern Recognition related clauses
+ */
+static void
+transformRPR(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+	/*
+	 * Window definition exists?
+	 */
+	if (windef == NULL)
+		return;
+
+	/*
+	 * Row Pattern Common Syntax clause exists?
+	 */
+	if (windef->rpCommonSyntax == NULL)
+		return;
+
+	/* Check Frame option. Frame must start at current row */
+	if ((wc->frameOptions & FRAMEOPTION_START_CURRENT_ROW) == 0)
+		ereport(ERROR,
+				(errcode(ERRCODE_SYNTAX_ERROR),
+				 errmsg("FRAME must start at current row when row patttern recognition is used")));
+
+	/* Transform AFTER MACH SKIP TO clause */
+	wc->rpSkipTo = windef->rpCommonSyntax->rpSkipTo;
+
+	/* Transform SEEK or INITIAL clause */
+	wc->initial = windef->rpCommonSyntax->initial;
+
+	/* Transform DEFINE clause into list of TargetEntry's */
+	wc->defineClause = transformDefineClause(pstate, wc, windef, targetlist);
+
+	/* Check PATTERN clause and copy to patternClause */
+	transformPatternClause(pstate, wc, windef);
+
+	/* Transform MEASURE clause */
+	transformMeasureClause(pstate, wc, windef);
+}
+
+/*
+ * transformDefineClause Process DEFINE clause and transform ResTarget into
+ *		list of TargetEntry.
+ *
+ * XXX we only support column reference in row pattern definition search
+ * condition, e.g. "price". <row pattern definition variable name>.<column
+ * reference> is not supported, e.g. "A.price".
+ */
+static List *
+transformDefineClause(ParseState *pstate, WindowClause *wc, WindowDef *windef, List **targetlist)
+{
+	/* DEFINE variable name initials */
+	static	char	*defineVariableInitials = "abcdefghijklmnopqrstuvwxyz";
+
+	ListCell		*lc, *l;
+	ResTarget		*restarget, *r;
+	List			*restargets;
+	char			*name;
+	int				initialLen;
+	int				i;
+
+	/*
+	 * If Row Definition Common Syntax exists, DEFINE clause must exist.
+	 * (the raw parser should have already checked it.)
+	 */
+	Assert(windef->rpCommonSyntax->rpDefs != NULL);
+
+	/*
+	 * Check and add "A AS A IS TRUE" if pattern variable is missing in DEFINE
+	 * per the SQL standard.
+	 */
+	restargets = NIL;
+	foreach(lc, windef->rpCommonSyntax->rpPatterns)
+	{
+		A_Expr	*a;
+		bool	found = false;
+
+		if (!IsA(lfirst(lc), A_Expr))
+			ereport(ERROR,
+					errmsg("node type is not A_Expr"));
+
+		a = (A_Expr *)lfirst(lc);
+		name = strVal(a->lexpr);
+
+		foreach(l, windef->rpCommonSyntax->rpDefs)
+		{
+			restarget = (ResTarget *)lfirst(l);
+
+			if (!strcmp(restarget->name, name))
+			{
+				found = true;
+				break;
+			}
+		}
+
+		if (!found)
+		{
+			/*
+			 * "name" is missing. So create "name AS name IS TRUE" ResTarget
+			 * node and add it to the temporary list.
+			 */
+			A_Const	   *n;
+
+			restarget = makeNode(ResTarget);
+			n = makeNode(A_Const);
+			n->val.boolval.type = T_Boolean;
+			n->val.boolval.boolval = true;
+			n->location = -1;
+			restarget->name = pstrdup(name);
+			restarget->indirection = NIL;
+			restarget->val = (Node *)n;
+			restarget->location = -1;
+			restargets = lappend((List *)restargets, restarget);
+		}
+	}
+
+	if (list_length(restargets) >= 1)
+	{
+		/* add missing DEFINEs */
+		windef->rpCommonSyntax->rpDefs = list_concat(windef->rpCommonSyntax->rpDefs,
+													 restargets);
+		list_free(restargets);
+	}
+
+	/*
+	 * Check for duplicate row pattern definition variables.  The standard
+	 * requires that no two row pattern definition variable names shall be
+	 * equivalent.
+	 */
+	restargets = NIL;
+	foreach(lc, windef->rpCommonSyntax->rpDefs)
+	{
+		restarget = (ResTarget *)lfirst(lc);
+		name = restarget->name;
+
+		/*
+		 * Add DEFINE expression (Restarget->val) to the targetlist as a
+		 * TargetEntry if it does not exist yet. Planner will add the column
+		 * ref var node to the outer plan's target list later on. This makes
+		 * DEFINE expression could access the outer tuple while evaluating
+		 * PATTERN.
+		 *
+		 * XXX: adding whole expressions of DEFINE to the plan.targetlist is
+		 * not so good, because it's not necessary to evalute the expression
+		 * in the target list while running the plan. We should extract the
+		 * var nodes only then add them to the plan.targetlist.
+		 */
+		findTargetlistEntrySQL99(pstate, (Node *)restarget->val, targetlist, EXPR_KIND_RPR_DEFINE);
+
+		/*
+		 * Make sure that the row pattern definition search condition is a
+		 * boolean expression.
+		 */
+		transformWhereClause(pstate, restarget->val,
+							 EXPR_KIND_RPR_DEFINE, "DEFINE");
+
+		foreach(l, restargets)
+		{
+			char		*n;
+
+			r = (ResTarget *) lfirst(l);
+			n = r->name;
+
+			if (!strcmp(n, name))
+				ereport(ERROR,
+						(errcode(ERRCODE_SYNTAX_ERROR),
+						 errmsg("row pattern definition variable name \"%s\" appears more than once in DEFINE clause",
+								name),
+						 parser_errposition(pstate, exprLocation((Node *)r))));
+		}
+		restargets = lappend(restargets, restarget);
+	}
+	list_free(restargets);
+
+	/*
+	 * Create list of row pattern DEFINE variable name's initial.
+	 * We assign [a-z] to them (up to 26 variable names are allowed).
+	 */
+	restargets = NIL;
+	i = 0;
+	initialLen = strlen(defineVariableInitials);
+
+	foreach(lc, windef->rpCommonSyntax->rpDefs)
+	{
+		char	initial[2];
+
+		restarget = (ResTarget *)lfirst(lc);
+		name = restarget->name;
+
+		if (i >= initialLen)
+		{
+			ereport(ERROR,
+					(errcode(ERRCODE_SYNTAX_ERROR),
+					 errmsg("number of row pattern definition variable names exceeds %d", initialLen),
+					 parser_errposition(pstate, exprLocation((Node *)restarget))));
+		}
+		initial[0] = defineVariableInitials[i++];
+		initial[1] = '\0';
+		wc->defineInitial = lappend(wc->defineInitial, makeString(pstrdup(initial)));
+	}
+
+	return transformTargetList(pstate, windef->rpCommonSyntax->rpDefs,
+							   EXPR_KIND_RPR_DEFINE);
+}
+
+/*
+ * transformPatternClause
+ *		Process PATTERN clause and return PATTERN clause in the raw parse tree
+ */
+static void
+transformPatternClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+	ListCell	*lc, *l;
+
+	/*
+	 * Row Pattern Common Syntax clause exists?
+	 */
+	if (windef->rpCommonSyntax == NULL)
+		return;
+
+	/*
+	 * Primary row pattern variable names in PATTERN clause must appear in
+	 * DEFINE clause as row pattern definition variable names.
+	 */
+	wc->patternVariable = NIL;
+	wc->patternRegexp = NIL;
+	foreach(lc, windef->rpCommonSyntax->rpPatterns)
+	{
+		A_Expr	*a;
+		char	*name;
+		char	*regexp;
+		bool	found = false;
+
+		if (!IsA(lfirst(lc), A_Expr))
+			ereport(ERROR,
+					errmsg("node type is not A_Expr"));
+
+		a = (A_Expr *)lfirst(lc);
+		name = strVal(a->lexpr);
+
+		foreach(l, windef->rpCommonSyntax->rpDefs)
+		{
+			ResTarget	*restarget = (ResTarget *)lfirst(l);
+
+			if (!strcmp(restarget->name, name))
+			{
+				found = true;
+				break;
+			}
+		}
+
+		if (!found)
+		{
+			ereport(ERROR,
+					(errcode(ERRCODE_SYNTAX_ERROR),
+					 errmsg("primary row pattern variable name \"%s\" does not appear in DEFINE clause",
+							name),
+					 parser_errposition(pstate, exprLocation((Node *)a))));
+		}
+		wc->patternVariable = lappend(wc->patternVariable, makeString(pstrdup(name)));
+		regexp = strVal(lfirst(list_head(a->name)));
+		wc->patternRegexp = lappend(wc->patternRegexp, makeString(pstrdup(regexp)));
+	}
+}
+
+/*
+ * transformMeasureClause
+ *		Process MEASURE clause
+ *	XXX MEASURE clause is not supported yet
+ */
+static List *
+transformMeasureClause(ParseState *pstate, WindowClause *wc, WindowDef *windef)
+{
+	if (windef->rowPatternMeasures == NIL)
+		return NIL;
+
+	ereport(ERROR,
+			(errcode(ERRCODE_SYNTAX_ERROR),
+			 errmsg("%s","MEASURE clause is not supported yet"),
+			 parser_errposition(pstate, exprLocation((Node *)windef->rowPatternMeasures))));
+}
diff --git a/src/backend/parser/parse_expr.c b/src/backend/parser/parse_expr.c
index 64c582c344..18b58ac263 100644
--- a/src/backend/parser/parse_expr.c
+++ b/src/backend/parser/parse_expr.c
@@ -557,6 +557,7 @@ transformColumnRef(ParseState *pstate, ColumnRef *cref)
 		case EXPR_KIND_COPY_WHERE:
 		case EXPR_KIND_GENERATED_COLUMN:
 		case EXPR_KIND_CYCLE_MARK:
+		case EXPR_KIND_RPR_DEFINE:
 			/* okay */
 			break;
 
@@ -1770,6 +1771,7 @@ transformSubLink(ParseState *pstate, SubLink *sublink)
 		case EXPR_KIND_VALUES:
 		case EXPR_KIND_VALUES_SINGLE:
 		case EXPR_KIND_CYCLE_MARK:
+		case EXPR_KIND_RPR_DEFINE:
 			/* okay */
 			break;
 		case EXPR_KIND_CHECK_CONSTRAINT:
@@ -3149,6 +3151,8 @@ ParseExprKindName(ParseExprKind exprKind)
 			return "GENERATED AS";
 		case EXPR_KIND_CYCLE_MARK:
 			return "CYCLE";
+		case EXPR_KIND_RPR_DEFINE:
+			return "DEFINE";
 
 			/*
 			 * There is intentionally no default: case here, so that the
diff --git a/src/backend/parser/parse_func.c b/src/backend/parser/parse_func.c
index b3f0b6a137..2ff3699538 100644
--- a/src/backend/parser/parse_func.c
+++ b/src/backend/parser/parse_func.c
@@ -2656,6 +2656,9 @@ check_srf_call_placement(ParseState *pstate, Node *last_srf, int location)
 		case EXPR_KIND_CYCLE_MARK:
 			errkind = true;
 			break;
+		case EXPR_KIND_RPR_DEFINE:
+			errkind = true;
+			break;
 
 			/*
 			 * There is intentionally no default: case here, so that the
-- 
2.25.1


----Next_Part(Sat_Sep__2_15_52_35_2023_273)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="v5-0003-Row-pattern-recognition-patch-planner.patch"



^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v2] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index fb6a4914b06..1da073f0aaa 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index d3febfe58f1..075194a8af8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1844,6 +1844,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index 6d304f32fb0..d346aefe9b6 100644
--- a/meson.build
+++ b/meson.build
@@ -2694,6 +2694,7 @@ decl_checks = [
   ['strlcpy', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index 563f20374ff..b41a5f9dcca 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#if HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#if HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#if HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 339268dc8ef..e7b8e023829 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -80,6 +80,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of `memset_s', and to 0 if you
    don't. */
 #undef HAVE_DECL_MEMSET_S

base-commit: 6831cd9e3b082d7b830c3196742dd49e3540c49b
-- 
2.49.0


--7u2oub7w3nl66bkx--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread

* [PATCH v1] Use open file description locks for data directory lockfile
@ 2025-12-18 17:21 Dmitrii Dolgov <9erthalion6@gmail.com>
  0 siblings, 0 replies; 659+ messages in thread

From: Dmitrii Dolgov @ 2025-12-18 17:21 UTC (permalink / raw)

When starting up, postmaster checks for an existing data directory lockfile. If
this file contains current process PID, it's assumed to be stale. Turns out
there is another possibility: we might be running in a PID namespace, and there
is another postgres running inside another PID namespace using the same data
directory. The result is that we don't see another process due to namespace
isolation and start concurrently with the other.

To prevent such situations, at startup use fcntl to get an exclusive open file
description lock for data directory lockfile. Since such locks are associated
with open file descriptors, meaning they're not affected by PID namespace
isolation. It's a "best effort" locking, intended to work with already existing
mechanism, not replace it.

This approach was discussed multiple times in the past, and usually was
rejected as the main work horse for the data directory lockfile due to:

* Portability issues. Open file description lock was a non-POSIX extension in
  Linux and similar flock is from BSD standard. But looks like everybody agrees
  that such locks make more sense than a typical advisory locks, and
  F_OFD_SETLK made its way into POSIX.1 2024 [1].

* Issues with NFS. The current state of things here looks like this:

  - NFSv3 doesn't implement open file description locks, they're converted to
    advisory locks instead. Advisory locks are subject to namespace isolation,
    meaning that processes in different PID namespaces will not see each other
    advisory lock, and it's still possible to run multiple postgres
    instances on the same data directory.

  - NFSv4 uses a lease system for locking, I haven't found any mention of
    conversion to advisory locks neither in the man page nor in RFC [2].

To summarize, the approach is now considered POSIX and should fix the described
problem everywhere, except NFSv3.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530
---
 configure                         |  14 ++++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 107 +++++++++++++++++++++++++-----
 src/include/pg_config.h.in        |   4 ++
 5 files changed, 111 insertions(+), 18 deletions(-)

diff --git a/configure b/configure
index 14ad0a5006f..b176ac39799 100755
--- a/configure
+++ b/configure
@@ -16177,6 +16177,20 @@ cat >>confdefs.h <<_ACEOF
 _ACEOF
 
 
+# Linux open file descriptor locks
+ac_fn_c_check_decl "$LINENO" "F_OFD_SETLK" "ac_cv_have_decl_F_OFD_SETLK" "#include <fcntl.h>
+"
+if test "x$ac_cv_have_decl_F_OFD_SETLK" = xyes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_F_OFD_SETLK $ac_have_decl
+_ACEOF
+
+
 ac_fn_c_check_func "$LINENO" "explicit_bzero" "ac_cv_func_explicit_bzero"
 if test "x$ac_cv_func_explicit_bzero" = xyes; then :
   $as_echo "#define HAVE_EXPLICIT_BZERO 1" >>confdefs.h
diff --git a/configure.ac b/configure.ac
index 01b3bbc1be8..d6cf1f27771 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1838,6 +1838,9 @@ AC_CHECK_DECLS([memset_s], [], [], [#define __STDC_WANT_LIB_EXT1__ 1
 # This is probably only present on macOS, but may as well check always
 AC_CHECK_DECLS(F_FULLFSYNC, [], [], [#include <fcntl.h>])
 
+# Linux open file descriptor locks
+AC_CHECK_DECLS([F_OFD_SETLK], [], [], [#include <fcntl.h>])
+
 AC_REPLACE_FUNCS(m4_normalize([
 	explicit_bzero
 	getopt
diff --git a/meson.build b/meson.build
index d7c5193d4ce..ad9ecad829f 100644
--- a/meson.build
+++ b/meson.build
@@ -2667,6 +2667,7 @@ decl_checks = [
   ['strnlen', 'string.h'],
   ['strsep',  'string.h'],
   ['timingsafe_bcmp',  'string.h'],
+  ['F_OFD_SETLK', 'fcntl.h'],
 ]
 
 # Need to check for function declarations for these functions, because
diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c
index fec79992c8d..78bb7df543e 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -68,6 +68,11 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+/* File descriptor for data directory lock file. */
+static int DataDirLockFD;
+#endif
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1117,6 +1122,45 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * Flock the data directory lockfile.
+ *
+ * Lock the data directory lockfile with an open file description lock. If the
+ * lock is already taken, it's a hard stop. It's only a best effort test, and
+ * any other errors are ignored. On succes the file descriptor is duplicated,
+ * to make sure there will be at least one open copy of it to keep the lock.
+ *
+ * filename is used only for reporting purposes.
+ */
+static void
+FlockDataDirLockFile(int fd, const char *filename)
+{
+
+#ifdef HAVE_DECL_F_OFD_SETLK
+	struct flock lock;
+
+	lock.l_type		= F_WRLCK;
+	lock.l_whence	= SEEK_SET;
+	lock.l_start	= 0;
+	lock.l_len		= 0;
+	lock.l_pid		= 0;
+
+	if (fcntl(fd, F_OFD_SETLK, &lock) == -1)
+	{
+		if (errno == EAGAIN)
+			ereport(FATAL,
+					(errcode(ERRCODE_LOCK_FILE_EXISTS),
+					 errmsg("cannot lock the lock file \"%s\"", filename),
+					 errhint("Another server is starting.")));
+		else
+			elog(WARNING, "Failed locking file \"%s\", %m", filename);
+	}
+	else
+		DataDirLockFD = dup(fd);
+#endif
+
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1125,6 +1169,11 @@ UnlinkLockFiles(int status, Datum arg)
 {
 	ListCell   *l;
 
+#ifdef HAVE_DECL_F_OFD_SETLK
+	/* Close the file descriptor, which keeps the open file description lock */
+	close(DataDirLockFD);
+#endif
+
 	foreach(l, lock_files)
 	{
 		char	   *curfile = (char *) lfirst(l);
@@ -1171,22 +1220,32 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	const char *envvar;
 
 	/*
-	 * If the PID in the lockfile is our own PID or our parent's or
-	 * grandparent's PID, then the file must be stale (probably left over from
-	 * a previous system boot cycle).  We need to check this because of the
-	 * likelihood that a reboot will assign exactly the same PID as we had in
-	 * the previous reboot, or one that's only one or two counts larger and
-	 * hence the lockfile's PID now refers to an ancestor shell process.  We
-	 * allow pg_ctl to pass down its parent shell PID (our grandparent PID)
-	 * via the environment variable PG_GRANDPARENT_PID; this is so that
-	 * launching the postmaster via pg_ctl can be just as reliable as
-	 * launching it directly.  There is no provision for detecting
-	 * further-removed ancestor processes, but if the init script is written
-	 * carefully then all but the immediate parent shell will be root-owned
-	 * processes and so the kill test will fail with EPERM.  Note that we
-	 * cannot get a false negative this way, because an existing postmaster
-	 * would surely never launch a competing postmaster or pg_ctl process
-	 * directly.
+	 * If we find an already existing lockfile containing our own PID,
+	 * there are few options:
+	 *
+	 * - There is another process, that we don't see due to PID namespace
+	 *   isolation, which is already running in this data directory.
+	 *
+	 *   To prevent two concurrent processes working with the same data
+	 *   directory, we first try to lock the lockfile exclusively.
+	 *
+	 * - The file must be stale, probably left over from a previous system boot
+	 *   cycle. The same if the lockfile contains our parent's or grandparent's
+	 *   PID.
+	 *
+	 *   We need to check this because of the likelihood that a reboot will
+	 *   assign exactly the same PID as we had in the previous reboot, or one
+	 *   that's only one or two counts larger and hence the lockfile's PID now
+	 *   refers to an ancestor shell process.  We allow pg_ctl to pass down its
+	 *   parent shell PID (our grandparent PID) via the environment variable
+	 *   PG_GRANDPARENT_PID; this is so that launching the postmaster via
+	 *   pg_ctl can be just as reliable as launching it directly.  There is no
+	 *   provision for detecting further-removed ancestor processes, but if the
+	 *   init script is written carefully then all but the immediate parent
+	 *   shell will be root-owned processes and so the kill test will fail with
+	 *   EPERM.  Note that we cannot get a false negative this way, because an
+	 *   existing postmaster would surely never launch a competing postmaster
+	 *   or pg_ctl process directly.
 	 */
 	my_pid = getpid();
 
@@ -1222,7 +1281,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		 */
 		fd = open(filename, O_RDWR | O_CREAT | O_EXCL, pg_file_create_mode);
 		if (fd >= 0)
-			break;				/* Success; exit the retry loop */
+		{
+			/* Success; lock and exit the retry loop */
+			FlockDataDirLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1236,8 +1299,12 @@ CreateLockFile(const char *filename, bool amPostmaster,
 		/*
 		 * Read the file to get the old owner's PID.  Note race condition
 		 * here: file might have been deleted since we tried to create it.
+		 *
+		 * We're going to use the same fd for flock, and want to create a write
+		 * lock for the latter one. Since both fd and the lock have to be of
+		 * the same type, open the file for read and write.
 		 */
-		fd = open(filename, O_RDONLY, pg_file_create_mode);
+		fd = open(filename, O_RDWR, pg_file_create_mode);
 		if (fd < 0)
 		{
 			if (errno == ENOENT)
@@ -1247,6 +1314,10 @@ CreateLockFile(const char *filename, bool amPostmaster,
 					 errmsg("could not open lock file \"%s\": %m",
 							filename)));
 		}
+
+		/* Try to lock the file. We stop here, if it's already locked. */
+		FlockDataDirLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 92fcc5f3063..c19a50f108e 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -83,6 +83,10 @@
    don't. */
 #undef HAVE_DECL_F_FULLFSYNC
 
+/* Define to 1 if you have the declaration of `F_OFD_SETLK', and to 0 if you
+   don't. */
+#undef HAVE_DECL_F_OFD_SETLK
+
 /* Define to 1 if you have the declaration of
    `LLVMCreateGDBRegistrationListener', and to 0 if you don't. */
 #undef HAVE_DECL_LLVMCREATEGDBREGISTRATIONLISTENER

base-commit: 44f49511b7940adf3be4337d4feb2de38fe92297
-- 
2.49.0


--n6t6hie3zne7u6vd--





^ permalink  raw  reply  [nested|flat] 659+ messages in thread