agora inbox for [email protected]  

help / color / mirror / Atom feed

[PATCH] Allow CREATE INDEX CONCURRENTLY on partitioned table
372+ messages / 2 participants
[nested] [flat]

* [PATCH] Allow CREATE INDEX CONCURRENTLY on partitioned table
@ 2020-06-06 22:42 Justin Pryzby <[email protected]>
  0 siblings, 0 replies; 372+ messages in thread

From: Justin Pryzby @ 2020-06-06 22:42 UTC (permalink / raw)

https://www.postgresql.org/message-id/flat/[email protected]
---
 doc/src/sgml/ddl.sgml                  |   4 +-
 doc/src/sgml/ref/create_index.sgml     |  14 +-
 src/backend/commands/indexcmds.c       | 200 ++++++++++++++++++-------
 src/test/regress/expected/indexing.out | 127 +++++++++++++++-
 src/test/regress/sql/indexing.sql      |  26 +++-
 5 files changed, 297 insertions(+), 74 deletions(-)

diff --git a/doc/src/sgml/ddl.sgml b/doc/src/sgml/ddl.sgml
index 91c036d1cbe..64efdf1e879 100644
--- a/doc/src/sgml/ddl.sgml
+++ b/doc/src/sgml/ddl.sgml
@@ -4178,9 +4178,7 @@ ALTER TABLE measurement ATTACH PARTITION measurement_y2008m02
      so that they are applied automatically to the entire hierarchy.
      This is very
      convenient, as not only will the existing partitions become indexed, but
-     also any partitions that are created in the future will.  One limitation is
-     that it's not possible to use the <literal>CONCURRENTLY</literal>
-     qualifier when creating such a partitioned index.  To avoid long lock
+     also any partitions that are created in the future will.  To avoid long lock
      times, it is possible to use <command>CREATE INDEX ON ONLY</command>
      the partitioned table; such an index is marked invalid, and the partitions
      do not get the index applied automatically.  The indexes on partitions can
diff --git a/doc/src/sgml/ref/create_index.sgml b/doc/src/sgml/ref/create_index.sgml
index 40986aa502f..b05102efdaf 100644
--- a/doc/src/sgml/ref/create_index.sgml
+++ b/doc/src/sgml/ref/create_index.sgml
@@ -645,7 +645,10 @@ CREATE [ UNIQUE ] INDEX [ CONCURRENTLY ] [ [ IF NOT EXISTS ] <replaceable class=
    <para>
     If a problem arises while scanning the table, such as a deadlock or a
     uniqueness violation in a unique index, the <command>CREATE INDEX</command>
-    command will fail but leave behind an <quote>invalid</quote> index. This index
+    command will fail but leave behind an <quote>invalid</quote> index.
+    If this happens while build an index concurrently on a partitioned
+    table, the command can also leave behind <quote>valid</quote> or
+    <quote>invalid</quote> indexes on table partitions.  The invalid index
     will be ignored for querying purposes because it might be incomplete;
     however it will still consume update overhead. The <application>psql</application>
     <command>\d</command> command will report such an index as <literal>INVALID</literal>:
@@ -692,15 +695,6 @@ Indexes:
     cannot.
    </para>
 
-   <para>
-    Concurrent builds for indexes on partitioned tables are currently not
-    supported.  However, you may concurrently build the index on each
-    partition individually and then finally create the partitioned index
-    non-concurrently in order to reduce the time where writes to the
-    partitioned table will be locked out.  In this case, building the
-    partitioned index is a metadata only operation.
-   </para>
-
   </refsect2>
  </refsect1>
 
diff --git a/src/backend/commands/indexcmds.c b/src/backend/commands/indexcmds.c
index 3ec8b5cca6c..daba8b67dbe 100644
--- a/src/backend/commands/indexcmds.c
+++ b/src/backend/commands/indexcmds.c
@@ -93,6 +93,11 @@ static char *ChooseIndexName(const char *tabname, Oid namespaceId,
 							 bool primary, bool isconstraint);
 static char *ChooseIndexNameAddition(List *colnames);
 static List *ChooseIndexColumnNames(List *indexElems);
+static void DefineIndexConcurrentInternal(Oid relationId,
+										  Oid indexRelationId,
+										  IndexInfo *indexInfo,
+										  LOCKTAG heaplocktag,
+										  LockRelId heaprelid);
 static void ReindexIndex(RangeVar *indexRelation, ReindexParams *params,
 						 bool isTopLevel);
 static void RangeVarCallbackForReindexIndex(const RangeVar *relation,
@@ -559,7 +564,6 @@ DefineIndex(Oid relationId,
 	bool		amissummarizing;
 	amoptions_function amoptions;
 	bool		partitioned;
-	bool		safe_index;
 	Datum		reloptions;
 	int16	   *coloptions;
 	IndexInfo  *indexInfo;
@@ -567,12 +571,10 @@ DefineIndex(Oid relationId,
 	bits16		constr_flags;
 	int			numberOfAttributes;
 	int			numberOfKeyAttributes;
-	TransactionId limitXmin;
 	ObjectAddress address;
 	LockRelId	heaprelid;
 	LOCKTAG		heaplocktag;
 	LOCKMODE	lockmode;
-	Snapshot	snapshot;
 	Oid			root_save_userid;
 	int			root_save_sec_context;
 	int			root_save_nestlevel;
@@ -705,17 +707,6 @@ DefineIndex(Oid relationId,
 	partitioned = rel->rd_rel->relkind == RELKIND_PARTITIONED_TABLE;
 	if (partitioned)
 	{
-		/*
-		 * Note: we check 'stmt->concurrent' rather than 'concurrent', so that
-		 * the error is thrown also for temporary tables.  Seems better to be
-		 * consistent, even though we could do it on temporary table because
-		 * we're not actually doing it concurrently.
-		 */
-		if (stmt->concurrent)
-			ereport(ERROR,
-					(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
-					 errmsg("cannot create index on partitioned table \"%s\" concurrently",
-							RelationGetRelationName(rel))));
 		if (stmt->excludeOpNames)
 			ereport(ERROR,
 					(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
@@ -1089,10 +1080,6 @@ DefineIndex(Oid relationId,
 		}
 	}
 
-	/* Is index safe for others to ignore?  See set_indexsafe_procflags() */
-	safe_index = indexInfo->ii_Expressions == NIL &&
-		indexInfo->ii_Predicate == NIL;
-
 	/*
 	 * Report index creation if appropriate (delay this till after most of the
 	 * error checks)
@@ -1157,6 +1144,11 @@ DefineIndex(Oid relationId,
 		if (pd->nparts != 0)
 			flags |= INDEX_CREATE_INVALID;
 	}
+	else if (concurrent && OidIsValid(parentIndexId))
+	{
+		/* If concurrent, initially build index partitions as "invalid" */
+		flags |= INDEX_CREATE_INVALID;
+	}
 
 	if (stmt->deferrable)
 		constr_flags |= INDEX_CONSTR_CREATE_DEFERRABLE;
@@ -1494,58 +1486,54 @@ DefineIndex(Oid relationId,
 			 * invalid, this is incorrect, so update our row to invalid too.
 			 */
 			if (invalidate_parent)
-			{
-				Relation	pg_index = table_open(IndexRelationId, RowExclusiveLock);
-				HeapTuple	tup,
-							newtup;
-
-				tup = SearchSysCache1(INDEXRELID,
-									  ObjectIdGetDatum(indexRelationId));
-				if (!HeapTupleIsValid(tup))
-					elog(ERROR, "cache lookup failed for index %u",
-						 indexRelationId);
-				newtup = heap_copytuple(tup);
-				((Form_pg_index) GETSTRUCT(newtup))->indisvalid = false;
-				CatalogTupleUpdate(pg_index, &tup->t_self, newtup);
-				ReleaseSysCache(tup);
-				table_close(pg_index, RowExclusiveLock);
-				heap_freetuple(newtup);
-			}
+				index_set_state_flags(indexRelationId, INDEX_DROP_CLEAR_VALID);
 		}
 
 		/*
 		 * Indexes on partitioned tables are not themselves built, so we're
-		 * done here.
+		 * done here in the non-concurrent case.
 		 */
-		AtEOXact_GUC(false, root_save_nestlevel);
-		SetUserIdAndSecContext(root_save_userid, root_save_sec_context);
-		table_close(rel, NoLock);
-		if (!OidIsValid(parentIndexId))
-			pgstat_progress_end_command();
-		else
+		if (!concurrent)
 		{
-			/* Update progress for an intermediate partitioned index itself */
-			pgstat_progress_incr_param(PROGRESS_CREATEIDX_PARTITIONS_DONE, 1);
-		}
+			AtEOXact_GUC(false, root_save_nestlevel);
+			SetUserIdAndSecContext(root_save_userid, root_save_sec_context);
+			table_close(rel, NoLock);
 
-		return address;
+			if (!OidIsValid(parentIndexId))
+				pgstat_progress_end_command();
+			else
+			{
+				/*
+				 * Update progress for an intermediate partitioned index
+				 * itself
+				 */
+				pgstat_progress_incr_param(PROGRESS_CREATEIDX_PARTITIONS_DONE, 1);
+			}
+
+			return address;
+		}
 	}
 
 	AtEOXact_GUC(false, root_save_nestlevel);
 	SetUserIdAndSecContext(root_save_userid, root_save_sec_context);
 
-	if (!concurrent)
+	/*
+	 * All done in the non-concurrent case, and when building catalog entries
+	 * of partitions for CIC.
+	 */
+	if (!concurrent || OidIsValid(parentIndexId))
 	{
-		/* Close the heap and we're done, in the non-concurrent case */
 		table_close(rel, NoLock);
 
 		/*
 		 * If this is the top-level index, the command is done overall;
-		 * otherwise, increment progress to report one child index is done.
+		 * otherwise (when being called recursively), increment progress to
+		 * report that one child index is done.  Except in the concurrent
+		 * (catalog-only) case, which is handled later.
 		 */
 		if (!OidIsValid(parentIndexId))
 			pgstat_progress_end_command();
-		else
+		else if (!concurrent)
 			pgstat_progress_incr_param(PROGRESS_CREATEIDX_PARTITIONS_DONE, 1);
 
 		return address;
@@ -1556,6 +1544,114 @@ DefineIndex(Oid relationId,
 	SET_LOCKTAG_RELATION(heaplocktag, heaprelid.dbId, heaprelid.relId);
 	table_close(rel, NoLock);
 
+	if (!partitioned)
+	{
+		/* CREATE INDEX CONCURRENTLY on a nonpartitioned table */
+		DefineIndexConcurrentInternal(relationId, indexRelationId,
+									  indexInfo, heaplocktag, heaprelid);
+		pgstat_progress_end_command();
+		return address;
+	}
+	else
+	{
+		/*
+		 * For CIC on a partitioned table, finish by building indexes on
+		 * partitions
+		 */
+
+		ListCell   *lc;
+		List	   *childs;
+		List	   *partitioned = NIL;
+		MemoryContext cic_context,
+					old_context;
+
+		/* Create special memory context for cross-transaction storage */
+		cic_context = AllocSetContextCreate(PortalContext,
+											"Create index concurrently",
+											ALLOCSET_DEFAULT_SIZES);
+
+		old_context = MemoryContextSwitchTo(cic_context);
+		childs = find_all_inheritors(indexRelationId, ShareLock, NULL);
+		MemoryContextSwitchTo(old_context);
+
+		foreach(lc, childs)
+		{
+			Oid			indrelid = lfirst_oid(lc);
+			Oid			tabrelid;
+			char		relkind;
+
+			/*
+			 * Pre-existing partitions which were ATTACHED were already
+			 * counted in the progress report.
+			 */
+			if (get_index_isvalid(indrelid))
+				continue;
+
+			/*
+			 * Partitioned indexes are counted in the progress report, but
+			 * don't need to be further processed.
+			 */
+			relkind = get_rel_relkind(indrelid);
+			if (!RELKIND_HAS_STORAGE(relkind))
+			{
+				/* The toplevel index doesn't count towards "partitions done" */
+				if (indrelid != indexRelationId)
+					pgstat_progress_incr_param(PROGRESS_CREATEIDX_PARTITIONS_DONE, 1);
+
+				/*
+				 * Build up a list of all the intermediate partitioned tables
+				 * which will later need to be set valid.
+				 */
+				old_context = MemoryContextSwitchTo(cic_context);
+				partitioned = lappend_oid(partitioned, indrelid);
+				MemoryContextSwitchTo(old_context);
+				continue;
+			}
+
+			rel = table_open(relationId, ShareUpdateExclusiveLock);
+			heaprelid = rel->rd_lockInfo.lockRelId;
+			table_close(rel, ShareUpdateExclusiveLock);
+			SET_LOCKTAG_RELATION(heaplocktag, heaprelid.dbId, heaprelid.relId);
+
+			/* Process each partition in a separate transaction */
+			tabrelid = IndexGetRelation(indrelid, false);
+			DefineIndexConcurrentInternal(tabrelid, indrelid, indexInfo,
+										  heaplocktag, heaprelid);
+
+			PushActiveSnapshot(GetTransactionSnapshot());
+			pgstat_progress_incr_param(PROGRESS_CREATEIDX_PARTITIONS_DONE, 1);
+		}
+
+		/* Set as valid all partitioned indexes, including the parent */
+		foreach(lc, partitioned)
+		{
+			Oid			indrelid = lfirst_oid(lc);
+
+			index_set_state_flags(indrelid, INDEX_CREATE_SET_READY);
+			CommandCounterIncrement();
+			index_set_state_flags(indrelid, INDEX_CREATE_SET_VALID);
+		}
+
+		MemoryContextDelete(cic_context);
+		pgstat_progress_end_command();
+		PopActiveSnapshot();
+		return address;
+	}
+}
+
+
+static void
+DefineIndexConcurrentInternal(Oid relationId,
+							  Oid indexRelationId, IndexInfo *indexInfo,
+							  LOCKTAG heaplocktag, LockRelId heaprelid)
+{
+	TransactionId limitXmin;
+	Snapshot	snapshot;
+
+	/* Is index safe for others to ignore?  See set_indexsafe_procflags() */
+	bool		safe_index = indexInfo->ii_Expressions == NIL &&
+		indexInfo->ii_Predicate == NIL;
+
 	/*
 	 * For a concurrent build, it's important to make the catalog entries
 	 * visible to other transactions before we start to build the index. That
@@ -1751,10 +1847,6 @@ DefineIndex(Oid relationId,
 	 * Last thing to do is release the session-level lock on the parent table.
 	 */
 	UnlockRelationIdForSession(&heaprelid, ShareUpdateExclusiveLock);
-
-	pgstat_progress_end_command();
-
-	return address;
 }
 
 
diff --git a/src/test/regress/expected/indexing.out b/src/test/regress/expected/indexing.out
index 1bdd430f063..f1beee6d240 100644
--- a/src/test/regress/expected/indexing.out
+++ b/src/test/regress/expected/indexing.out
@@ -50,11 +50,130 @@ select relname, relkind, relhassubclass, inhparent::regclass
 (8 rows)
 
 drop table idxpart;
--- Some unsupported features
+-- CIC on partitioned table
 create table idxpart (a int, b int, c text) partition by range (a);
-create table idxpart1 partition of idxpart for values from (0) to (10);
-create index concurrently on idxpart (a);
-ERROR:  cannot create index on partitioned table "idxpart" concurrently
+create table idxpart1 partition of idxpart for values from (0) to (10) partition by range(a);
+create table idxpart11 partition of idxpart1 for values from (0) to (10) partition by range(a);
+create table idxpart111 partition of idxpart11 default partition by range(a);
+create table idxpart1111 partition of idxpart111 default partition by range(a);
+create table idxpart2 partition of idxpart for values from (10) to (20);
+create table idxpart3 partition of idxpart for values from (30) to (40) partition by range(a);
+create table idxpart31 partition of idxpart3 default;
+insert into idxpart2 values(10),(10); -- not unique
+create index concurrently on idxpart11 (a); -- partitioned and partition, with no leaves
+create index concurrently on idxpart1 (a); -- partitioned and partition
+create index concurrently on idxpart2 (a); -- leaf
+create index concurrently on idxpart (a); -- partitioned
+create unique index concurrently on idxpart (a); -- partitioned, unique failure
+ERROR:  could not create unique index "idxpart2_a_idx1"
+DETAIL:  Key (a)=(10) is duplicated.
+\d idxpart
+        Partitioned table "public.idxpart"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition key: RANGE (a)
+Indexes:
+    "idxpart_a_idx" btree (a)
+    "idxpart_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 3 (Use \d+ to list them.)
+
+\d idxpart1
+        Partitioned table "public.idxpart1"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart FOR VALUES FROM (0) TO (10)
+Partition key: RANGE (a)
+Indexes:
+    "idxpart1_a_idx" btree (a)
+    "idxpart1_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 1 (Use \d+ to list them.)
+
+\d idxpart11
+       Partitioned table "public.idxpart11"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart1 FOR VALUES FROM (0) TO (10)
+Partition key: RANGE (a)
+Indexes:
+    "idxpart11_a_idx" btree (a)
+    "idxpart11_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 1 (Use \d+ to list them.)
+
+\d idxpart111
+       Partitioned table "public.idxpart111"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart11 DEFAULT
+Partition key: RANGE (a)
+Indexes:
+    "idxpart111_a_idx" btree (a)
+    "idxpart111_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 1 (Use \d+ to list them.)
+
+\d idxpart1111
+      Partitioned table "public.idxpart1111"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart111 DEFAULT
+Partition key: RANGE (a)
+Indexes:
+    "idxpart1111_a_idx" btree (a)
+    "idxpart1111_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 0
+
+\d idxpart2
+              Table "public.idxpart2"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart FOR VALUES FROM (10) TO (20)
+Indexes:
+    "idxpart2_a_idx" btree (a)
+    "idxpart2_a_idx1" UNIQUE, btree (a) INVALID
+
+\d idxpart3
+        Partitioned table "public.idxpart3"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart FOR VALUES FROM (30) TO (40)
+Partition key: RANGE (a)
+Indexes:
+    "idxpart3_a_idx" btree (a)
+    "idxpart3_a_idx1" UNIQUE, btree (a) INVALID
+Number of partitions: 1 (Use \d+ to list them.)
+
+\d idxpart31
+             Table "public.idxpart31"
+ Column |  Type   | Collation | Nullable | Default 
+--------+---------+-----------+----------+---------
+ a      | integer |           |          | 
+ b      | integer |           |          | 
+ c      | text    |           |          | 
+Partition of: idxpart3 DEFAULT
+Indexes:
+    "idxpart31_a_idx" btree (a)
+    "idxpart31_a_idx1" UNIQUE, btree (a) INVALID
+
 drop table idxpart;
 -- Verify bugfix with query on indexed partitioned table with no partitions
 -- https://postgr.es/m/[email protected]
diff --git a/src/test/regress/sql/indexing.sql b/src/test/regress/sql/indexing.sql
index 429120e7104..fb0baedcc28 100644
--- a/src/test/regress/sql/indexing.sql
+++ b/src/test/regress/sql/indexing.sql
@@ -29,10 +29,30 @@ select relname, relkind, relhassubclass, inhparent::regclass
 	where relname like 'idxpart%' order by relname;
 drop table idxpart;
 
--- Some unsupported features
+-- CIC on partitioned table
 create table idxpart (a int, b int, c text) partition by range (a);
-create table idxpart1 partition of idxpart for values from (0) to (10);
-create index concurrently on idxpart (a);
+create table idxpart1 partition of idxpart for values from (0) to (10) partition by range(a);
+create table idxpart11 partition of idxpart1 for values from (0) to (10) partition by range(a);
+create table idxpart111 partition of idxpart11 default partition by range(a);
+create table idxpart1111 partition of idxpart111 default partition by range(a);
+create table idxpart2 partition of idxpart for values from (10) to (20);
+create table idxpart3 partition of idxpart for values from (30) to (40) partition by range(a);
+create table idxpart31 partition of idxpart3 default;
+
+insert into idxpart2 values(10),(10); -- not unique
+create index concurrently on idxpart11 (a); -- partitioned and partition, with no leaves
+create index concurrently on idxpart1 (a); -- partitioned and partition
+create index concurrently on idxpart2 (a); -- leaf
+create index concurrently on idxpart (a); -- partitioned
+create unique index concurrently on idxpart (a); -- partitioned, unique failure
+\d idxpart
+\d idxpart1
+\d idxpart11
+\d idxpart111
+\d idxpart1111
+\d idxpart2
+\d idxpart3
+\d idxpart31
 drop table idxpart;
 
 -- Verify bugfix with query on indexed partitioned table with no partitions
-- 
2.34.1


--naDu/P2UbiQYQWIS--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v4] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 35b0b72f0a7..25ebcd3cc47 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 0e624fe36b9..677137207e7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 d88a7a70308..153fbb477bb 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..d4c2f80eb46 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+}			PGLockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1;
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	PGLockFile *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(PGLockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		PGLockFile *lock_file = (PGLockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index ffb413ab612..ad34142e7d6 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1956,6 +1956,7 @@ PGIOAlignedBlock
 PGLZ_HistEntry
 PGLZ_Strategy
 PGLoadBalanceType
+PGLockFile
 PGMessageField
 PGModuleMagicFunction
 PGNoticeHooks

base-commit: 73dfe79fd6034b1e7e41e83d9c82c166dba8eb67
-- 
2.52.0


--tks2nr37zts5e6h7--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread

* [PATCH v3] Use open file description locks for lockfiles
@ 2025-12-18 17:21 Dmitrii Dolgov <[email protected]>
  0 siblings, 0 replies; 372+ 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.

Use open file description lock for both data directory and socker
lockfiles, since both are affected in the same way.

[1]: https://pubs.opengroup.org/onlinepubs/9799919799/functions/fcntl.html
[2]: https://www.rfc-editor.org/rfc/rfc7530

Reviewed-by: Ilmar Yunusov <[email protected]>
---
 configure                         |  14 +++
 configure.ac                      |   3 +
 meson.build                       |   1 +
 src/backend/utils/init/miscinit.c | 136 ++++++++++++++++++++++++------
 src/include/pg_config.h.in        |   4 +
 src/tools/pgindent/typedefs.list  |   1 +
 6 files changed, 134 insertions(+), 25 deletions(-)

diff --git a/configure b/configure
index 5f77f3cac29..15bda6c6413 100755
--- a/configure
+++ b/configure
@@ -16444,6 +16444,20 @@ 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 61cee42daa7..e24cb7a6f01 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1913,6 +1913,9 @@ 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 568e0e150bf..3d641dc0403 100644
--- a/meson.build
+++ b/meson.build
@@ -2901,6 +2901,7 @@ 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 7ffc808073a..26c3324542c 100644
--- a/src/backend/utils/init/miscinit.c
+++ b/src/backend/utils/init/miscinit.c
@@ -69,6 +69,15 @@ static List *lock_files = NIL;
 
 static Latch LocalLatchData;
 
+typedef struct
+{
+	/* LockFile name. */
+	const char *filename;
+
+	/* File descriptor for open file description lock. */
+	int			fd;
+} LockFile;
+
 /* ----------------------------------------------------------------
  *		ignoring system indexes support stuff
  *
@@ -1119,6 +1128,48 @@ RestoreClientConnectionInfo(char *conninfo)
  *-------------------------------------------------------------------------
  */
 
+/*
+ * OFD lock the specified lockfile.
+ *
+ * Lock the 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 argument is used only for reporting purposes.
+ */
+static int
+OFDLockFile(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);
+			return -1;
+		}
+	}
+	else
+		return dup(fd);
+#else
+	return -1
+#endif
+}
+
 /*
  * proc_exit callback to remove lockfiles.
  */
@@ -1129,9 +1180,16 @@ UnlinkLockFiles(int status, Datum arg)
 
 	foreach(l, lock_files)
 	{
-		char	   *curfile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
-		unlink(curfile);
+		/*
+		 * Close the file descriptor, which keeps the open file description
+		 * lock.
+		 */
+		if (lock_file->fd > 0)
+			close(lock_file->fd);
+
+		unlink(lock_file->filename);
 		/* Should we complain if the unlink fails? */
 	}
 	/* Since we're about to exit, no need to reclaim storage */
@@ -1161,7 +1219,9 @@ CreateLockFile(const char *filename, bool amPostmaster,
 			   const char *socketDir,
 			   bool isDDLock, const char *refName)
 {
-	int			fd;
+	int			fd,
+				flock_fd = -1;
+	LockFile   *lock_file;
 	char		buffer[MAXPGPATH * 2 + 256];
 	int			ntries;
 	int			len;
@@ -1173,22 +1233,32 @@ 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();
 
@@ -1224,7 +1294,11 @@ 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 */
+			flock_fd = OFDLockFile(fd, filename);
+			break;
+		}
 
 		/*
 		 * Couldn't create the pid file. Probably it already exists.
@@ -1238,8 +1312,12 @@ 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)
@@ -1249,6 +1327,10 @@ 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. */
+		flock_fd = OFDLockFile(fd, filename);
+
 		pgstat_report_wait_start(WAIT_EVENT_LOCK_FILE_CREATE_READ);
 		if ((len = read(fd, buffer, sizeof(buffer) - 1)) < 0)
 			ereport(FATAL,
@@ -1448,7 +1530,11 @@ CreateLockFile(const char *filename, bool amPostmaster,
 	 * Use lcons so that the lock files are unlinked in reverse order of
 	 * creation; this is critical!
 	 */
-	lock_files = lcons(pstrdup(filename), lock_files);
+	lock_file = palloc0_object(LockFile);
+	lock_file->filename = pstrdup(filename);
+	lock_file->fd = flock_fd;
+
+	lock_files = lcons(lock_file, lock_files);
 }
 
 /*
@@ -1495,14 +1581,14 @@ TouchSocketLockFiles(void)
 
 	foreach(l, lock_files)
 	{
-		char	   *socketLockFile = (char *) lfirst(l);
+		LockFile   *lock_file = (LockFile *) lfirst(l);
 
 		/* No need to touch the data directory lock file, we trust */
-		if (strcmp(socketLockFile, DIRECTORY_LOCK_FILE) == 0)
+		if (strcmp(lock_file->filename, DIRECTORY_LOCK_FILE) == 0)
 			continue;
 
 		/* we just ignore any error here */
-		(void) utime(socketLockFile, NULL);
+		(void) utime(lock_file->filename, NULL);
 	}
 }
 
diff --git a/src/include/pg_config.h.in b/src/include/pg_config.h.in
index 4f8113c144b..cc38c06dc13 100644
--- a/src/include/pg_config.h.in
+++ b/src/include/pg_config.h.in
@@ -85,6 +85,10 @@
    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
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 1969d467c1d..185d69b5520 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1653,6 +1653,7 @@ LocationLen
 LockAcquireResult
 LockClauseStrength
 LockData
+LockFile
 LockInfoData
 LockInstanceData
 LockMethod

base-commit: 031904048aa22e7c70dc8e9c170e2743f9b0f090
-- 
2.52.0


--zdel3ow7bygx53fm--





^ permalink  raw  reply  [nested|flat] 372+ messages in thread