agora inbox for [email protected]
help / color / mirror / Atom feed[PATCH 2/4] demote: support demoting instance from production to standby
2+ messages / 2 participants
[nested] [flat]
* [PATCH 2/4] demote: support demoting instance from production to standby
@ 2020-04-10 16:01 Jehan-Guillaume de Rorthais <[email protected]>
0 siblings, 0 replies; 2+ messages in thread
From: Jehan-Guillaume de Rorthais @ 2020-04-10 16:01 UTC (permalink / raw)
This patch adds the ability for an in production read write
instance to step back as a read only standby. Two different
modes are supported: fast or smart. Fast demote mode cancel
transactions holding a xid to demote as fast as possible.
Smart demote mode wait for existing transactions holding a
xid to finish and forbid new ones.
The demote process is triggered by creating a "demote" or
"demote_fast" trigger file then signaling postmaster with
SIGUSR1. This patch adds "pg_ctl [-m {fast|smart}] demote"
and an SQL admin function will be added in a separate commit.
The demote procedure starts be setting the postmaster state
machine to PM_DEMOTING which ends when all backends are
read-only and set LocalXLogInsert=0 to forbid new writes.
When PM_DEMOTING finish, the state is set to PM_SHUTDOWN
to create a shutdown checkpoint and exit the checkpointer.
We could have kept the checkpointer around with some more
effort but there's no good reason worthing the additionnal
code complexity.
ShutdownXLOG now takes a boolean arg to handle demote
differently than a normal shutdown. This allows the
checkpointer to leave walsenders alive to demote faster.
Moreover, this might be useful in futur for eg.
implementing a controlled switchover over the replication
protocol.
The checkpointer set the cluster state as DB_DEMOTING so
the Startup process can detect the demote procedure in
StartupXLOG() and handle subsystems accordingly.
The startup process is started on checkpointer exit. The
postmaster state machine is then switched to PM_STARTUP.
Some sub-processes are kept alive during the demote procedure:
the stat collector, bgwriter and optionally archiver and wal
senders.
At the end of demote, USR1 is sent to all backends and wal
senders to set their environment as in recovery and cascading.
Discuss/Todo:
* add doc
* do not handle backup in progress during demote
* investigate snapshots shmem needs/init during recovery compare to
production
* cancel running checkpoint during demote
* replace with a END_OF_PRODUCTION xlog record?
---
src/backend/access/transam/twophase.c | 98 ++++++++
src/backend/access/transam/xlog.c | 320 ++++++++++++++++--------
src/backend/postmaster/checkpointer.c | 28 +++
src/backend/postmaster/pgstat.c | 3 +
src/backend/postmaster/postmaster.c | 235 ++++++++++++++++-
src/backend/storage/ipc/procarray.c | 2 +
src/backend/storage/ipc/procsignal.c | 30 +++
src/backend/storage/lmgr/lock.c | 12 +
src/backend/tcop/postgres.c | 44 ++++
src/backend/utils/init/globals.c | 1 +
src/bin/pg_controldata/pg_controldata.c | 2 +
src/bin/pg_ctl/pg_ctl.c | 117 +++++++++
src/include/access/twophase.h | 1 +
src/include/access/xlog.h | 23 +-
src/include/catalog/pg_control.h | 1 +
src/include/libpq/libpq-be.h | 2 +-
src/include/miscadmin.h | 1 +
src/include/pgstat.h | 1 +
src/include/postmaster/bgwriter.h | 1 +
src/include/storage/lock.h | 2 +
src/include/storage/procsignal.h | 4 +
src/include/tcop/tcopprot.h | 2 +
src/include/utils/pidfile.h | 1 +
23 files changed, 802 insertions(+), 129 deletions(-)
diff --git a/src/backend/access/transam/twophase.c b/src/backend/access/transam/twophase.c
index ef4f9981e3..31deb487ed 100644
--- a/src/backend/access/transam/twophase.c
+++ b/src/backend/access/transam/twophase.c
@@ -1557,6 +1557,104 @@ FinishPreparedTransaction(const char *gid, bool isCommit)
pfree(buf);
}
+/*
+ * ShutdownPreparedTransactions: clean prepared from sheared memory
+ *
+ * This is called during the demote process to clean the shared memory
+ * before the startup process load everything back in correctly
+ * for the standby mode.
+ *
+ * Note: this function assue all prepared transaction have been
+ * written to disk. In consequence, it must be called AFTER the demote
+ * shutdown checkpoint.
+ *
+ * FIXME demote: pay attention to the previous note when removing shutdown
+ * checkpoint from the demote procedure.
+ */
+void
+ShutdownPreparedTransactions(void)
+{
+ int i;
+
+ for (i = 0; i < TwoPhaseState->numPrepXacts; i++)
+ {
+ GlobalTransaction gxact;
+ PGPROC *proc;
+ TransactionId xid;
+ char *buf;
+ char *bufptr;
+ TwoPhaseFileHeader *hdr;
+ TransactionId latestXid;
+ TransactionId *children;
+
+ gxact = TwoPhaseState->prepXacts[i];
+ proc = &ProcGlobal->allProcs[gxact->pgprocno];
+ xid = gxact->xid;
+
+ /* Read and validate 2PC state data */
+ Assert(gxact->ondisk);
+ buf = ReadTwoPhaseFile(xid, false);
+
+ /*
+ * Disassemble the header area
+ */
+ hdr = (TwoPhaseFileHeader *) buf;
+ Assert(TransactionIdEquals(hdr->xid, xid));
+ bufptr = buf + MAXALIGN(sizeof(TwoPhaseFileHeader))
+ + MAXALIGN(hdr->gidlen);
+ children = (TransactionId *) bufptr;
+ bufptr += MAXALIGN(hdr->nsubxacts * sizeof(TransactionId))
+ + MAXALIGN(hdr->ncommitrels * sizeof(RelFileNode))
+ + MAXALIGN(hdr->nabortrels * sizeof(RelFileNode))
+ + MAXALIGN(hdr->ninvalmsgs * sizeof(SharedInvalidationMessage));
+
+ /* compute latestXid among all children */
+ latestXid = TransactionIdLatest(xid, hdr->nsubxacts, children);
+
+ /* remove dummy proc associated to the gaxt */
+ ProcArrayRemove(proc, latestXid);
+
+ /*
+ * This lock is probably not needed during the demote process
+ * as all backends are already gone.
+ */
+ LWLockAcquire(TwoPhaseStateLock, LW_EXCLUSIVE);
+
+ /* cleanup locks */
+ for (;;)
+ {
+ TwoPhaseRecordOnDisk *record = (TwoPhaseRecordOnDisk *) bufptr;
+
+ Assert(record->rmid <= TWOPHASE_RM_MAX_ID);
+ if (record->rmid == TWOPHASE_RM_END_ID)
+ break;
+
+ bufptr += MAXALIGN(sizeof(TwoPhaseRecordOnDisk));
+
+ if (record->rmid == TWOPHASE_RM_LOCK_ID)
+ lock_twophase_shutdown(xid, record->info,
+ (void *) bufptr, record->len);
+
+ bufptr += MAXALIGN(record->len);
+ }
+
+ /* and put it back in the freelist */
+ gxact->next = TwoPhaseState->freeGXacts;
+ TwoPhaseState->freeGXacts = gxact;
+
+ /*
+ * Release the lock as all callbacks are called and shared memory cleanup
+ * is done.
+ */
+ LWLockRelease(TwoPhaseStateLock);
+
+ pfree(buf);
+ }
+
+ TwoPhaseState->numPrepXacts -= i;
+ Assert(TwoPhaseState->numPrepXacts == 0);
+}
+
/*
* Scan 2PC state data in memory and call the indicated callbacks for each 2PC record.
*/
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index c0d79f192c..abc88ec241 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -6298,6 +6298,11 @@ CheckRequiredParameterValues(void)
/*
* This must be called ONCE during postmaster or standalone-backend startup
*/
+/*
+ * FIXME demote: part of the code here assume there's no other active
+ * processes before signal PMSIGNAL_RECOVERY_STARTED is sent.
+ */
+
void
StartupXLOG(void)
{
@@ -6321,6 +6326,7 @@ StartupXLOG(void)
XLogPageReadPrivate private;
bool promoted = false;
struct stat st;
+ bool is_demoting = false;
/*
* We should have an aux process resource owner to use, and we should not
@@ -6385,6 +6391,25 @@ StartupXLOG(void)
str_time(ControlFile->time))));
break;
+ case DB_DEMOTING:
+ ereport(LOG,
+ (errmsg("database system was demoted at %s",
+ str_time(ControlFile->time))));
+ is_demoting = true;
+ bgwriterLaunched = true;
+ InArchiveRecovery = true;
+ StandbyMode = true;
+
+ /*
+ * previous state was RECOVERY_STATE_DONE. We need to
+ * reinit it to something else so RecoveryInProgress()
+ * doesn't return false.
+ */
+ SpinLockAcquire(&XLogCtl->info_lck);
+ XLogCtl->SharedRecoveryState = RECOVERY_STATE_ARCHIVE;
+ SpinLockRelease(&XLogCtl->info_lck);
+ break;
+
default:
ereport(FATAL,
(errmsg("control file contains invalid database cluster state")));
@@ -6418,7 +6443,8 @@ StartupXLOG(void)
* persisted. To avoid that, fsync the entire data directory.
*/
if (ControlFile->state != DB_SHUTDOWNED &&
- ControlFile->state != DB_SHUTDOWNED_IN_RECOVERY)
+ ControlFile->state != DB_SHUTDOWNED_IN_RECOVERY &&
+ ControlFile->state != DB_DEMOTING)
{
RemoveTempXlogFiles();
SyncDataDirectory();
@@ -6674,7 +6700,8 @@ StartupXLOG(void)
(errmsg("could not locate a valid checkpoint record")));
}
memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- wasShutdown = ((record->xl_info & ~XLR_INFO_MASK) == XLOG_CHECKPOINT_SHUTDOWN);
+ wasShutdown = ((record->xl_info & ~XLR_INFO_MASK) == XLOG_CHECKPOINT_SHUTDOWN) &&
+ !is_demoting;
}
/*
@@ -6736,9 +6763,9 @@ StartupXLOG(void)
LastRec = RecPtr = checkPointLoc;
ereport(DEBUG1,
- (errmsg_internal("redo record is at %X/%X; shutdown %s",
+ (errmsg_internal("redo record is at %X/%X; %s checkpoint",
(uint32) (checkPoint.redo >> 32), (uint32) checkPoint.redo,
- wasShutdown ? "true" : "false")));
+ wasShutdown ? "shutdown" : is_demoting? "demote": "")));
ereport(DEBUG1,
(errmsg_internal("next transaction ID: " UINT64_FORMAT "; next OID: %u",
U64FromFullTransactionId(checkPoint.nextXid),
@@ -6772,47 +6799,7 @@ StartupXLOG(void)
checkPoint.newestCommitTsXid);
XLogCtl->ckptFullXid = checkPoint.nextXid;
- /*
- * Initialize replication slots, before there's a chance to remove
- * required resources.
- */
- StartupReplicationSlots();
-
- /*
- * Startup logical state, needs to be setup now so we have proper data
- * during crash recovery.
- */
- StartupReorderBuffer();
- /*
- * Startup MultiXact. We need to do this early to be able to replay
- * truncations.
- */
- StartupMultiXact();
-
- /*
- * Ditto for commit timestamps. Activate the facility if the setting is
- * enabled in the control file, as there should be no tracking of commit
- * timestamps done when the setting was disabled. This facility can be
- * started or stopped when replaying a XLOG_PARAMETER_CHANGE record.
- */
- if (ControlFile->track_commit_timestamp)
- StartupCommitTs();
-
- /*
- * Recover knowledge about replay progress of known replication partners.
- */
- StartupReplicationOrigin();
-
- /*
- * Initialize unlogged LSN. On a clean shutdown, it's restored from the
- * control file. On recovery, all unlogged relations are blown away, so
- * the unlogged LSN counter can be reset too.
- */
- if (ControlFile->state == DB_SHUTDOWNED)
- XLogCtl->unloggedLSN = ControlFile->unloggedLSN;
- else
- XLogCtl->unloggedLSN = FirstNormalUnloggedLSN;
/*
* We must replay WAL entries using the same TimeLineID they were created
@@ -6821,19 +6808,64 @@ StartupXLOG(void)
*/
ThisTimeLineID = checkPoint.ThisTimeLineID;
- /*
- * Copy any missing timeline history files between 'now' and the recovery
- * target timeline from archive to pg_wal. While we don't need those files
- * ourselves - the history file of the recovery target timeline covers all
- * the previous timelines in the history too - a cascading standby server
- * might be interested in them. Or, if you archive the WAL from this
- * server to a different archive than the primary, it'd be good for all the
- * history files to get archived there after failover, so that you can use
- * one of the old timelines as a PITR target. Timeline history files are
- * small, so it's better to copy them unnecessarily than not copy them and
- * regret later.
- */
- restoreTimeLineHistoryFiles(ThisTimeLineID, recoveryTargetTLI);
+ if (!is_demoting)
+ {
+ /*
+ * Initialize replication slots, before there's a chance to remove
+ * required resources.
+ */
+ StartupReplicationSlots();
+
+ /*
+ * Startup logical state, needs to be setup now so we have proper data
+ * during crash recovery.
+ */
+ StartupReorderBuffer();
+
+ /*
+ * Startup MultiXact. We need to do this early to be able to replay
+ * truncations.
+ */
+ StartupMultiXact();
+
+ /*
+ * Ditto for commit timestamps. Activate the facility if the setting is
+ * enabled in the control file, as there should be no tracking of commit
+ * timestamps done when the setting was disabled. This facility can be
+ * started or stopped when replaying a XLOG_PARAMETER_CHANGE record.
+ */
+ if (ControlFile->track_commit_timestamp)
+ StartupCommitTs();
+
+ /*
+ * Recover knowledge about replay progress of known replication partners.
+ */
+ StartupReplicationOrigin();
+
+ /*
+ * Initialize unlogged LSN. On a clean shutdown, it's restored from the
+ * control file. On recovery, all unlogged relations are blown away, so
+ * the unlogged LSN counter can be reset too.
+ */
+ if (ControlFile->state == DB_SHUTDOWNED)
+ XLogCtl->unloggedLSN = ControlFile->unloggedLSN;
+ else
+ XLogCtl->unloggedLSN = FirstNormalUnloggedLSN;
+
+ /*
+ * Copy any missing timeline history files between 'now' and the recovery
+ * target timeline from archive to pg_wal. While we don't need those files
+ * ourselves - the history file of the recovery target timeline covers all
+ * the previous timelines in the history too - a cascading standby server
+ * might be interested in them. Or, if you archive the WAL from this
+ * server to a different archive than the master, it'd be good for all the
+ * history files to get archived there after failover, so that you can use
+ * one of the old timelines as a PITR target. Timeline history files are
+ * small, so it's better to copy them unnecessarily than not copy them and
+ * regret later.
+ */
+ restoreTimeLineHistoryFiles(ThisTimeLineID, recoveryTargetTLI);
+ }
/*
* Before running in recovery, scan pg_twophase and fill in its status to
@@ -6888,11 +6920,25 @@ StartupXLOG(void)
dbstate_at_startup = ControlFile->state;
if (InArchiveRecovery)
{
- ControlFile->state = DB_IN_ARCHIVE_RECOVERY;
+ if (is_demoting)
+ {
+ /*
+ * Avoid concurrent access to the ControlFile datas
+ * during demotion.
+ */
+ LWLockAcquire(ControlFileLock, LW_EXCLUSIVE);
+ ControlFile->state = DB_IN_ARCHIVE_RECOVERY;
+ LWLockRelease(ControlFileLock);
+ }
+ else
+ {
+ ControlFile->state = DB_IN_ARCHIVE_RECOVERY;
- SpinLockAcquire(&XLogCtl->info_lck);
- XLogCtl->SharedRecoveryState = RECOVERY_STATE_ARCHIVE;
- SpinLockRelease(&XLogCtl->info_lck);
+ /* This is already set if demoting */
+ SpinLockAcquire(&XLogCtl->info_lck);
+ XLogCtl->SharedRecoveryState = RECOVERY_STATE_ARCHIVE;
+ SpinLockRelease(&XLogCtl->info_lck);
+ }
}
else
{
@@ -6982,7 +7028,8 @@ StartupXLOG(void)
/*
* Reset pgstat data, because it may be invalid after recovery.
*/
- pgstat_reset_all();
+ if (!is_demoting)
+ pgstat_reset_all();
/*
* If there was a backup label file, it's done its job and the info
@@ -7044,7 +7091,7 @@ StartupXLOG(void)
InitRecoveryTransactionEnvironment();
- if (wasShutdown)
+ if (wasShutdown || is_demoting)
oldestActiveXID = PrescanPreparedTransactions(&xids, &nxids);
else
oldestActiveXID = checkPoint.oldestActiveXid;
@@ -7057,6 +7104,11 @@ StartupXLOG(void)
* Startup commit log and subtrans only. MultiXact and commit
* timestamp have already been started up and other SLRUs are not
* maintained during recovery and need not be started yet.
+ *
+ * Starting up commit log is technicaly not needed during demote
+ * as the in-memory data did not move. However, this is a
+ * lightweight initialization and this might seem expected as
+ * pure symmetry as ShutdownCLOG() is called during ShutdownXLog().
*/
StartupCLOG();
StartupSUBTRANS(oldestActiveXID);
@@ -7067,7 +7119,7 @@ StartupXLOG(void)
* empty running-xacts record and use that here and now. Recover
* additional standby state for prepared transactions.
*/
- if (wasShutdown)
+ if (wasShutdown || is_demoting)
{
RunningTransactionsData running;
TransactionId latestCompletedXid;
@@ -7938,6 +7990,7 @@ StartupXLOG(void)
SpinLockAcquire(&XLogCtl->info_lck);
XLogCtl->SharedRecoveryState = RECOVERY_STATE_DONE;
+ XLogCtl->SharedHotStandbyActive = false;
SpinLockRelease(&XLogCtl->info_lck);
UpdateControlFile();
@@ -8056,6 +8109,23 @@ CheckRecoveryConsistency(void)
}
}
+/*
+ * Initialize the local TimeLineID
+ */
+bool
+SetLocalRecoveryInProgress(void)
+{
+ /*
+ * use volatile pointer to make sure we make a fresh read of the
+ * shared variable.
+ */
+ volatile XLogCtlData *xlogctl = XLogCtl;
+
+ LocalRecoveryInProgress = (xlogctl->SharedRecoveryState != RECOVERY_STATE_DONE);
+
+ return LocalRecoveryInProgress;
+}
+
/*
* Is the system still in recovery?
*
@@ -8077,13 +8147,7 @@ RecoveryInProgress(void)
return false;
else
{
- /*
- * use volatile pointer to make sure we make a fresh read of the
- * shared variable.
- */
- volatile XLogCtlData *xlogctl = XLogCtl;
-
- LocalRecoveryInProgress = (xlogctl->SharedRecoveryState != RECOVERY_STATE_DONE);
+ SetLocalRecoveryInProgress();
/*
* Initialize TimeLineID and RedoRecPtr when we discover that recovery
@@ -8503,6 +8567,8 @@ GetLastSegSwitchData(XLogRecPtr *lastSwitchLSN)
void
ShutdownXLOG(int code, Datum arg)
{
+ bool is_demoting = DatumGetBool(arg);
+
/*
* We should have an aux process resource owner to use, and we should not
* be in a transaction that's installed some other resowner.
@@ -8512,35 +8578,60 @@ ShutdownXLOG(int code, Datum arg)
CurrentResourceOwner == AuxProcessResourceOwner);
CurrentResourceOwner = AuxProcessResourceOwner;
- /* Don't be chatty in standalone mode */
- ereport(IsPostmasterEnvironment ? LOG : NOTICE,
- (errmsg("shutting down")));
-
- /*
- * Signal walsenders to move to stopping state.
- */
- WalSndInitStopping();
-
- /*
- * Wait for WAL senders to be in stopping state. This prevents commands
- * from writing new WAL.
- */
- WalSndWaitStopping();
+ if (is_demoting)
+ {
+ /*
+ * In contrast with normal shutdown, we keep wal senders alive
+ * during demote. First, this allows the demote to complete faster.
+ * Second, we might need them in futur to implement a controlled
+ * switchover over the replication protocol.
+ */
+ /* Don't be chatty in standalone mode */
+ ereport(IsPostmasterEnvironment ? LOG : NOTICE,
+ (errmsg("demoting")));
- if (RecoveryInProgress())
- CreateRestartPoint(CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_IMMEDIATE);
+ /*
+ * FIXME demote: avoiding checkpoint?
+ * A checkpoint is probably running during a demote action. If
+ * we don't want to wait for the checkpoint during the demote,
+ * we might need to cancel it as it will not be able to write
+ * to the WAL after the demote.
+ */
+ CreateCheckPoint(CHECKPOINT_IS_DEMOTE | CHECKPOINT_IMMEDIATE);
+ ShutdownPreparedTransactions();
+ }
else
{
+ /* Don't be chatty in standalone mode */
+ ereport(IsPostmasterEnvironment ? LOG : NOTICE,
+ (errmsg("shutting down")));
+
/*
- * If archiving is enabled, rotate the last XLOG file so that all the
- * remaining records are archived (postmaster wakes up the archiver
- * process one more time at the end of shutdown). The checkpoint
- * record will go to the next XLOG file and won't be archived (yet).
+ * Signal walsenders to move to stopping state.
*/
- if (XLogArchivingActive() && XLogArchiveCommandSet())
- RequestXLogSwitch(false);
+ WalSndInitStopping();
+
+ /*
+ * Wait for WAL senders to be in stopping state. This prevents commands
+ * from writing new WAL.
+ */
+ WalSndWaitStopping();
+
+ if (RecoveryInProgress())
+ CreateRestartPoint(CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_IMMEDIATE);
+ else
+ {
+ /*
+ * If archiving is enabled, rotate the last XLOG file so that all the
+ * remaining records are archived (postmaster wakes up the archiver
+ * process one more time at the end of shutdown). The checkpoint
+ * record will go to the next XLOG file and won't be archived (yet).
+ */
+ if (XLogArchivingActive() && XLogArchiveCommandSet())
+ RequestXLogSwitch(false);
- CreateCheckPoint(CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_IMMEDIATE);
+ CreateCheckPoint(CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_IMMEDIATE);
+ }
}
ShutdownCLOG();
ShutdownCommitTs();
@@ -8554,9 +8645,10 @@ ShutdownXLOG(int code, Datum arg)
static void
LogCheckpointStart(int flags, bool restartpoint)
{
- elog(LOG, "%s starting:%s%s%s%s%s%s%s%s",
+ elog(LOG, "%s starting:%s%s%s%s%s%s%s%s%s",
restartpoint ? "restartpoint" : "checkpoint",
(flags & CHECKPOINT_IS_SHUTDOWN) ? " shutdown" : "",
+ (flags & CHECKPOINT_IS_DEMOTE) ? " demote" : "",
(flags & CHECKPOINT_END_OF_RECOVERY) ? " end-of-recovery" : "",
(flags & CHECKPOINT_IMMEDIATE) ? " immediate" : "",
(flags & CHECKPOINT_FORCE) ? " force" : "",
@@ -8692,6 +8784,7 @@ UpdateCheckPointDistanceEstimate(uint64 nbytes)
*
* flags is a bitwise OR of the following:
* CHECKPOINT_IS_SHUTDOWN: checkpoint is for database shutdown.
+ * CHECKPOINT_IS_DEMOTE: checkpoint is for demote.
* CHECKPOINT_END_OF_RECOVERY: checkpoint is for end of WAL recovery.
* CHECKPOINT_IMMEDIATE: finish the checkpoint ASAP,
* ignoring checkpoint_completion_target parameter.
@@ -8720,6 +8813,7 @@ void
CreateCheckPoint(int flags)
{
bool shutdown;
+ bool demote;
CheckPoint checkPoint;
XLogRecPtr recptr;
XLogSegNo _logSegNo;
@@ -8732,14 +8826,21 @@ CreateCheckPoint(int flags)
int nvxids;
/*
- * An end-of-recovery checkpoint is really a shutdown checkpoint, just
- * issued at a different time.
+ * An end-of-recovery or demote checkpoint is really a shutdown checkpoint,
+ * just issued at a different time.
*/
- if (flags & (CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_END_OF_RECOVERY))
+ if (flags & (CHECKPOINT_IS_SHUTDOWN |
+ CHECKPOINT_IS_DEMOTE |
+ CHECKPOINT_END_OF_RECOVERY))
shutdown = true;
else
shutdown = false;
+ if (flags & CHECKPOINT_IS_DEMOTE)
+ demote = true;
+ else
+ demote = false;
+
/* sanity check */
if (RecoveryInProgress() && (flags & CHECKPOINT_END_OF_RECOVERY) == 0)
elog(ERROR, "can't create a checkpoint during recovery");
@@ -8780,7 +8881,7 @@ CreateCheckPoint(int flags)
if (shutdown)
{
LWLockAcquire(ControlFileLock, LW_EXCLUSIVE);
- ControlFile->state = DB_SHUTDOWNING;
+ ControlFile->state = demote? DB_DEMOTING:DB_SHUTDOWNING;
ControlFile->time = (pg_time_t) time(NULL);
UpdateControlFile();
LWLockRelease(ControlFileLock);
@@ -8826,7 +8927,7 @@ CreateCheckPoint(int flags)
* avoid inserting duplicate checkpoints when the system is idle.
*/
if ((flags & (CHECKPOINT_IS_SHUTDOWN | CHECKPOINT_END_OF_RECOVERY |
- CHECKPOINT_FORCE)) == 0)
+ CHECKPOINT_IS_DEMOTE | CHECKPOINT_FORCE)) == 0)
{
if (last_important_lsn == ControlFile->checkPoint)
{
@@ -8994,8 +9095,8 @@ CreateCheckPoint(int flags)
* allows us to reconstruct the state of running transactions during
* archive recovery, if required. Skip, if this info disabled.
*
- * If we are shutting down, or Startup process is completing crash
- * recovery we don't need to write running xact data.
+ * If we are shutting down, demoting or Startup process is completing
+ * crash recovery we don't need to write running xact data.
*/
if (!shutdown && XLogStandbyInfoActive())
LogStandbySnapshot();
@@ -9014,11 +9115,11 @@ CreateCheckPoint(int flags)
XLogFlush(recptr);
/*
- * We mustn't write any new WAL after a shutdown checkpoint, or it will be
- * overwritten at next startup. No-one should even try, this just allows
- * sanity-checking. In the case of an end-of-recovery checkpoint, we want
- * to just temporarily disable writing until the system has exited
- * recovery.
+ * We mustn't write any new WAL after a shutdown or demote checkpoint, or
+ * it will be overwritten at next startup. No-one should even try, this
+ * just allows sanity-checking. In the case of an end-of-recovery
+ * checkpoint, we want to just temporarily disable writing until the system
+ * has exited recovery.
*/
if (shutdown)
{
@@ -9034,7 +9135,8 @@ CreateCheckPoint(int flags)
*/
if (shutdown && checkPoint.redo != ProcLastRecPtr)
ereport(PANIC,
- (errmsg("concurrent write-ahead log activity while database system is shutting down")));
+ (errmsg("concurrent write-ahead log activity while database system is %s",
+ demote? "demoting":"shutting down")));
/*
* Remember the prior checkpoint's redo ptr for
@@ -9047,7 +9149,7 @@ CreateCheckPoint(int flags)
*/
LWLockAcquire(ControlFileLock, LW_EXCLUSIVE);
if (shutdown)
- ControlFile->state = DB_SHUTDOWNED;
+ ControlFile->state = demote? DB_DEMOTING:DB_SHUTDOWNED;
ControlFile->checkPoint = ProcLastRecPtr;
ControlFile->checkPointCopy = checkPoint;
ControlFile->time = (pg_time_t) time(NULL);
diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c
index 624a3238b8..58473a61fd 100644
--- a/src/backend/postmaster/checkpointer.c
+++ b/src/backend/postmaster/checkpointer.c
@@ -151,6 +151,7 @@ double CheckPointCompletionTarget = 0.5;
* Private state
*/
static bool ckpt_active = false;
+static volatile sig_atomic_t demoteRequestPending = false;
/* these values are valid when ckpt_active is true: */
static pg_time_t ckpt_start_time;
@@ -552,6 +553,21 @@ HandleCheckpointerInterrupts(void)
*/
UpdateSharedMemoryConfig();
}
+ if (demoteRequestPending)
+ {
+ demoteRequestPending = false;
+ /* Close down the database */
+ ShutdownXLOG(0, BoolGetDatum(true));
+ /*
+ * Exit checkpointer. We could keep it around during demotion, but
+ * exiting here has multiple benefices:
+ * - to create a fresh process with clean local vars
+ * (eg. LocalRecoveryInProgress)
+ * - to signal postmaster the demote shutdown checkpoint is done
+ * and keep going with next steps of the demotion
+ */
+ proc_exit(0);
+ }
if (ShutdownRequestPending)
{
/*
@@ -680,6 +696,7 @@ CheckpointWriteDelay(int flags, double progress)
* in which case we just try to catch up as quickly as possible.
*/
if (!(flags & CHECKPOINT_IMMEDIATE) &&
+ !demoteRequestPending &&
!ShutdownRequestPending &&
!ImmediateCheckpointRequested() &&
IsCheckpointOnSchedule(progress))
@@ -812,6 +829,17 @@ IsCheckpointOnSchedule(double progress)
* --------------------------------
*/
+/* SIGUSR1: set flag to demote */
+void
+ReqCheckpointDemoteHandler(SIGNAL_ARGS)
+{
+ int save_errno = errno;
+
+ demoteRequestPending = true;
+
+ errno = save_errno;
+}
+
/* SIGINT: set flag to run a normal checkpoint right away */
static void
ReqCheckpointHandler(SIGNAL_ARGS)
diff --git a/src/backend/postmaster/pgstat.c b/src/backend/postmaster/pgstat.c
index 73ce944fb1..e75d33d335 100644
--- a/src/backend/postmaster/pgstat.c
+++ b/src/backend/postmaster/pgstat.c
@@ -3854,6 +3854,9 @@ pgstat_get_wait_ipc(WaitEventIPC w)
case WAIT_EVENT_PROMOTE:
event_name = "Promote";
break;
+ case WAIT_EVENT_DEMOTE:
+ event_name = "Demote";
+ break;
case WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT:
event_name = "RecoveryConflictSnapshot";
break;
diff --git a/src/backend/postmaster/postmaster.c b/src/backend/postmaster/postmaster.c
index 42223c0f61..231febaf2f 100644
--- a/src/backend/postmaster/postmaster.c
+++ b/src/backend/postmaster/postmaster.c
@@ -277,12 +277,13 @@ static StartupStatusEnum StartupStatus = STARTUP_NOT_RUNNING;
#define ImmediateShutdown 3
static int Shutdown = NoShutdown;
+static bool DemoteSignal = false; /* true on demote request */
static bool FatalError = false; /* T if recovering from backend crash */
/*
- * We use a simple state machine to control startup, shutdown, and
- * crash recovery (which is rather like shutdown followed by startup).
+ * We use a simple state machine to control startup, shutdown, demote and
+ * crash recovery (both are rather like shutdown followed by startup).
*
* After doing all the postmaster initialization work, we enter PM_STARTUP
* state and the startup process is launched. The startup process begins by
@@ -325,6 +326,7 @@ typedef enum
{
PM_INIT, /* postmaster starting */
PM_STARTUP, /* waiting for startup subprocess */
+ PM_DEMOTING, /* waiting for idle or RO backends for demote */
PM_RECOVERY, /* in archive recovery mode */
PM_HOT_STANDBY, /* in hot standby mode */
PM_RUN, /* normal "database is alive" state */
@@ -429,10 +431,14 @@ static bool RandomCancelKey(int32 *cancel_key);
static void signal_child(pid_t pid, int signal);
static bool SignalSomeChildren(int signal, int targets);
static void TerminateChildren(int signal);
+static void RemoveDemoteSignalFiles(void);
+static bool CheckDemoteSignal(void);
+
#define SignalChildren(sig) SignalSomeChildren(sig, BACKEND_TYPE_ALL)
static int CountChildren(int target);
+static int CountXacts(void);
static bool assign_backendlist_entry(RegisteredBgWorker *rw);
static void maybe_start_bgworkers(void);
static bool CreateOptsFile(int argc, char *argv[], char *fullprogname);
@@ -2319,6 +2325,11 @@ retry1:
(errcode(ERRCODE_CANNOT_CONNECT_NOW),
errmsg("the database system is starting up")));
break;
+ case CAC_DEMOTE:
+ ereport(FATAL,
+ (errcode(ERRCODE_CANNOT_CONNECT_NOW),
+ errmsg("the database system is demoting")));
+ break;
case CAC_SHUTDOWN:
ereport(FATAL,
(errcode(ERRCODE_CANNOT_CONNECT_NOW),
@@ -2450,16 +2461,19 @@ canAcceptConnections(int backend_type)
CAC_state result = CAC_OK;
/*
- * Can't start backends when in startup/shutdown/inconsistent recovery
- * state. We treat autovac workers the same as user backends for this
- * purpose. However, bgworkers are excluded from this test; we expect
- * bgworker_should_start_now() decided whether the DB state allows them.
+ * Can't start backends when in startup/demote/shutdown/inconsistent
+ * recovery state. We treat autovac workers the same as user backends
+ * for this purpose. However, bgworkers are excluded from this test; we
+ * expect bgworker_should_start_now() decided whether the DB state allows
+ * them.
*/
if (pmState != PM_RUN && pmState != PM_HOT_STANDBY &&
backend_type != BACKEND_TYPE_BGWORKER)
{
if (Shutdown > NoShutdown)
return CAC_SHUTDOWN; /* shutdown is pending */
+ else if (DemoteSignal)
+ return CAC_DEMOTE; /* demote is pending */
else if (!FatalError &&
(pmState == PM_STARTUP ||
pmState == PM_RECOVERY))
@@ -3091,7 +3105,18 @@ reaper(SIGNAL_ARGS)
if (pid == CheckpointerPID)
{
CheckpointerPID = 0;
- if (EXIT_STATUS_0(exitstatus) && pmState == PM_SHUTDOWN)
+ if (EXIT_STATUS_0(exitstatus) &&
+ DemoteSignal &&
+ pmState == PM_SHUTDOWN)
+ {
+ /*
+ * The checkpointer exit signals the demote shutdown checkpoint
+ * is done. The startup recovery mode can be started from there.
+ */
+ ereport(DEBUG1,
+ (errmsg_internal("checkpointer shutdown for demote")));
+ }
+ else if (EXIT_STATUS_0(exitstatus) && pmState == PM_SHUTDOWN)
{
/*
* OK, we saw normal exit of the checkpointer after it's been
@@ -3799,6 +3824,25 @@ PostmasterStateMachine(void)
}
}
+ if (pmState == PM_DEMOTING)
+ {
+ int numXacts = CountXacts();
+
+ /*
+ * PM_DEMOTING state ends when we have no active transactions
+ * and all backends set LocalXLogInsertAllowed=0
+ */
+ if (numXacts == 0)
+ {
+ ereport(LOG, (errmsg("all backends in read only")));
+
+ SendProcSignal(CheckpointerPID, PROCSIG_CHECKPOINTER_DEMOTING, InvalidBackendId);
+ pmState = PM_SHUTDOWN;
+ }
+ else
+ ereport(LOG, (errmsg("waiting for %d transactions to finish", numXacts)));
+ }
+
/*
* If we're ready to do so, signal child processes to shut down. (This
* isn't a persistent state, but treating it as a distinct pmState allows
@@ -4002,6 +4046,20 @@ PostmasterStateMachine(void)
(StartupStatus == STARTUP_CRASHED || !restart_after_crash))
ExitPostmaster(1);
+
+ /* Demoting: start the Startup Process */
+ if (DemoteSignal && pmState == PM_SHUTDOWN && CheckpointerPID == 0)
+ {
+ /* stop archiver process if not required during standby */
+ if (!XLogArchivingAlways() && PgArchPID != 0)
+ signal_child(PgArchPID, SIGQUIT);
+
+ StartupPID = StartupDataBase();
+ Assert(StartupPID != 0);
+ StartupStatus = STARTUP_RUNNING;
+ pmState = PM_STARTUP;
+ }
+
/*
* If we need to recover from a crash, wait for all non-syslogger children
* to exit, then reset shmem and StartupDataBase.
@@ -5212,8 +5270,12 @@ sigusr1_handler(SIGNAL_ARGS)
* Crank up the background tasks. It doesn't matter if this fails,
* we'll just try again later.
*/
+ if (!DemoteSignal)
+ Assert(PgArchPID == 0);
+
Assert(CheckpointerPID == 0);
CheckpointerPID = StartCheckpointer();
+
Assert(BgWriterPID == 0);
BgWriterPID = StartBackgroundWriter();
@@ -5221,8 +5283,7 @@ sigusr1_handler(SIGNAL_ARGS)
* Start the archiver if we're responsible for (re-)archiving received
* files.
*/
- Assert(PgArchPID == 0);
- if (XLogArchivingAlways())
+ if (PgArchPID == 0 && XLogArchivingAlways())
PgArchPID = pgarch_start();
/*
@@ -5233,6 +5294,7 @@ sigusr1_handler(SIGNAL_ARGS)
if (!EnableHotStandby)
{
AddToDataDirLockFile(LOCK_FILE_LINE_PM_STATUS, PM_STATUS_STANDBY);
+ DemoteSignal = false;
#ifdef USE_SYSTEMD
sd_notify(0, "READY=1");
#endif
@@ -5243,11 +5305,15 @@ sigusr1_handler(SIGNAL_ARGS)
if (CheckPostmasterSignal(PMSIGNAL_BEGIN_HOT_STANDBY) &&
pmState == PM_RECOVERY && Shutdown == NoShutdown)
{
+ dlist_iter iter;
+
/*
* Likewise, start other special children as needed.
*/
- Assert(PgStatPID == 0);
- PgStatPID = pgstat_start();
+ if (!DemoteSignal)
+ Assert(PgStatPID == 0);
+ if(PgStatPID == 0)
+ PgStatPID = pgstat_start();
ereport(LOG,
(errmsg("database system is ready to accept read only connections")));
@@ -5258,8 +5324,18 @@ sigusr1_handler(SIGNAL_ARGS)
sd_notify(0, "READY=1");
#endif
+ if (DemoteSignal)
+ dlist_foreach(iter, &BackendList)
+ {
+ Backend *bp = dlist_container(Backend, elem, iter.cur);
+
+ if (!bp->dead_end && bp->bkend_type & (BACKEND_TYPE_NORMAL|BACKEND_TYPE_WALSND))
+ SendProcSignal(bp->pid, PROCSIG_DEMOTED, InvalidBackendId);
+ }
+
pmState = PM_HOT_STANDBY;
connsAllowed = ALLOW_ALL_CONNS;
+ DemoteSignal = false;
/* Some workers may be scheduled to start now */
StartWorkerNeeded = true;
@@ -5351,6 +5427,97 @@ sigusr1_handler(SIGNAL_ARGS)
signal_child(StartupPID, SIGUSR2);
}
+ if (CheckDemoteSignal() && pmState != PM_RUN )
+ {
+ DemoteSignal = false;
+ RemoveDemoteSignalFiles();
+ ereport(LOG,
+ (errmsg("ignoring demote signal because already in standby mode")));
+ }
+ /* received demote signal */
+ else if (CheckDemoteSignal())
+ {
+ FILE *standby_file;
+ dlist_iter iter;
+ bool fast_demote;
+ struct stat stat_buf;
+
+ fast_demote = (stat(DEMOTE_FAST_SIGNAL_FILE, &stat_buf) == 0);
+
+ DemoteSignal = true;
+ RemoveDemoteSignalFiles();
+
+ /* create the standby signal file */
+ standby_file = AllocateFile(STANDBY_SIGNAL_FILE, "w");
+ if (!standby_file)
+ {
+ ereport(ERROR,
+ (errcode_for_file_access(),
+ errmsg("could not create file \"%s\": %m",
+ STANDBY_SIGNAL_FILE)));
+ goto out;
+ }
+
+ if (FreeFile(standby_file))
+ {
+ ereport(ERROR,
+ (errcode_for_file_access(),
+ errmsg("could not write file \"%s\": %m",
+ STANDBY_SIGNAL_FILE)));
+ unlink(STANDBY_SIGNAL_FILE);
+ goto out;
+ }
+
+ if (fast_demote == 0)
+ {
+ /* smart demote */
+ ereport(LOG, (errmsg("received smart demote request")));
+
+ }
+ else
+ {
+ /* fast demote */
+ ereport(LOG, (errmsg("received fast demote request")));
+ }
+
+ SignalSomeChildren(SIGTERM,
+ BACKEND_TYPE_AUTOVAC | BACKEND_TYPE_BGWORKER);
+
+ /* and the autovac launcher too */
+ if (AutoVacPID != 0)
+ signal_child(AutoVacPID, SIGTERM);
+ /* and the bgwriter too */
+ if (BgWriterPID != 0)
+ signal_child(BgWriterPID, SIGTERM);
+ /* and the walwriter too */
+ if (WalWriterPID != 0)
+ signal_child(WalWriterPID, SIGTERM);
+
+ dlist_foreach(iter, &BackendList)
+ {
+ Backend *bp = dlist_container(Backend, elem, iter.cur);
+
+ if (bp->dead_end)
+ continue;
+ /*
+ * Assign bkend_type for any recently announced WAL Sender
+ * processes.
+ */
+ if (bp->bkend_type == BACKEND_TYPE_NORMAL &&
+ ! IsPostmasterChildWalSender(bp->child_slot))
+ SendProcSignal(bp->pid,
+ (fast_demote?PROCSIG_DEMOTING_FAST:PROCSIG_DEMOTING),
+ InvalidBackendId);
+ }
+
+ pmState = PM_DEMOTING;
+
+ /* Report status */
+ AddToDataDirLockFile(LOCK_FILE_LINE_PM_STATUS, PM_STATUS_DEMOTING);
+ }
+
+out:
+
#ifdef WIN32
PG_SETMASK(&UnBlockSig);
#endif
@@ -5448,6 +5615,26 @@ CountChildren(int target)
}
+/*
+ * Count up the number of active transactions
+ */
+static int
+CountXacts(void)
+{
+ int i;
+ int cnt = 0;
+
+ for (i = 0; i < ProcGlobal->allProcCount; ++i)
+ {
+ PGPROC *proc = &ProcGlobal->allProcs[i];
+ if (TransactionIdIsValid(proc->xid))
+ cnt++;
+ }
+
+ return cnt;
+}
+
+
/*
* StartChildProcess -- start an auxiliary process for the postmaster
*
@@ -5912,6 +6099,7 @@ bgworker_should_start_now(BgWorkerStartTime start_time)
case PM_SHUTDOWN:
case PM_WAIT_BACKENDS:
case PM_STOP_BACKENDS:
+ case PM_DEMOTING:
break;
case PM_RUN:
@@ -6660,3 +6848,28 @@ InitPostmasterDeathWatchHandle(void)
GetLastError())));
#endif /* WIN32 */
}
+
+/*
+ * Remove the files signaling a demote request.
+ */
+static void
+RemoveDemoteSignalFiles(void)
+{
+ unlink(DEMOTE_SIGNAL_FILE);
+ unlink(DEMOTE_FAST_SIGNAL_FILE);
+}
+
+/*
+ * Check if a demote request appeared.
+ */
+static bool
+CheckDemoteSignal(void)
+{
+ struct stat stat_buf;
+
+ if (stat(DEMOTE_SIGNAL_FILE, &stat_buf) == 0 ||
+ stat(DEMOTE_FAST_SIGNAL_FILE, &stat_buf) == 0)
+ return true;
+
+ return false;
+}
diff --git a/src/backend/storage/ipc/procarray.c b/src/backend/storage/ipc/procarray.c
index 96e4a87857..83d7e05944 100644
--- a/src/backend/storage/ipc/procarray.c
+++ b/src/backend/storage/ipc/procarray.c
@@ -349,6 +349,8 @@ ProcArrayShmemSize(void)
size = add_size(size, mul_size(sizeof(int), PROCARRAY_MAXPROCS));
/*
+ * FIXME demote: check safe hotStandby related init and snapshot mech.
+ *
* During Hot Standby processing we have a data structure called
* KnownAssignedXids, created in shared memory. Local data structures are
* also created in various backends during GetSnapshotData(),
diff --git a/src/backend/storage/ipc/procsignal.c b/src/backend/storage/ipc/procsignal.c
index 4fa385b0ec..ac14c662d3 100644
--- a/src/backend/storage/ipc/procsignal.c
+++ b/src/backend/storage/ipc/procsignal.c
@@ -28,6 +28,7 @@
#include "storage/shmem.h"
#include "storage/sinval.h"
#include "tcop/tcopprot.h"
+#include "postmaster/bgwriter.h"
/*
* The SIGUSR1 signal is multiplexed to support signaling multiple event
@@ -585,6 +586,35 @@ procsignal_sigusr1_handler(SIGNAL_ARGS)
if (CheckProcSignal(PROCSIG_RECOVERY_CONFLICT_BUFFERPIN))
RecoveryConflictInterrupt(PROCSIG_RECOVERY_CONFLICT_BUFFERPIN);
+ /* signal checkpoint process to ignite a demote procedure */
+ if (CheckProcSignal(PROCSIG_CHECKPOINTER_DEMOTING))
+ ReqCheckpointDemoteHandler(PROCSIG_CHECKPOINTER_DEMOTING);
+
+ /*
+ * ask backends to enter in read only by setting
+ * LocalXLogInsertAllowed = 0 as soon as their active xact
+ * finished
+ */
+ if (CheckProcSignal(PROCSIG_DEMOTING))
+ ReqDemoteHandler(PROCSIG_DEMOTING);
+
+ /*
+ * ask backends to enter in read only by setting
+ * LocalXLogInsertAllowed = 0 if they are idle, or
+ * interrupt their current xact and terminate.
+ */
+ if (CheckProcSignal(PROCSIG_DEMOTING_FAST))
+ ReqDemoteHandler(PROCSIG_DEMOTING_FAST);
+
+ /*
+ * demote complete. Ask beckends to rely on
+ * recovery status for LocalXLogInsertAllowed by
+ * setting it to -1.
+ * WAL sender set am_cascading.
+ */
+ if (CheckProcSignal(PROCSIG_DEMOTED))
+ ReqDemotedHandler(PROCSIG_DEMOTED);
+
SetLatch(MyLatch);
latch_sigusr1_handler();
diff --git a/src/backend/storage/lmgr/lock.c b/src/backend/storage/lmgr/lock.c
index d86566f455..5843db0991 100644
--- a/src/backend/storage/lmgr/lock.c
+++ b/src/backend/storage/lmgr/lock.c
@@ -4370,6 +4370,18 @@ lock_twophase_postabort(TransactionId xid, uint16 info,
lock_twophase_postcommit(xid, info, recdata, len);
}
+/*
+ * 2PC shutdown from lock table.
+ *
+ * This is actually just the same as the COMMIT case.
+ */
+void
+lock_twophase_shutdown(TransactionId xid, uint16 info,
+ void *recdata, uint32 len)
+{
+ lock_twophase_postcommit(xid, info, recdata, len);
+}
+
/*
* VirtualXactLockTableInsert
*
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
index c9424f167c..b44e6f5876 100644
--- a/src/backend/tcop/postgres.c
+++ b/src/backend/tcop/postgres.c
@@ -67,6 +67,7 @@
#include "rewrite/rewriteHandler.h"
#include "storage/bufmgr.h"
#include "storage/ipc.h"
+#include "storage/pmsignal.h"
#include "storage/proc.h"
#include "storage/procsignal.h"
#include "storage/sinval.h"
@@ -3211,6 +3212,42 @@ ProcessInterrupts(void)
HandleParallelMessages();
}
+/* SIGUSR1: set flag to demote */
+void
+ReqDemoteHandler(ProcSignalReason reason)
+{
+ if (MyBackendType != B_BACKEND)
+ return;
+
+ if (TransactionIdIsValid(MyProc->xid))
+ {
+ if (reason == PROCSIG_DEMOTING_FAST)
+ {
+ InterruptPending = true;
+ ProcDiePending = true;
+ SetLatch(MyLatch);
+ }
+ else
+ DemotePending = true;
+ }
+ else
+ LocalSetXLogInsertNotAllowed();
+}
+
+/* SIGUSR1: reset LocalRecoveryInProgress */
+void
+ReqDemotedHandler(ProcSignalReason reason)
+{
+ ereport(LOG,
+ (errmsg("received demote complete signal")));
+
+ SetLocalRecoveryInProgress();
+ LocalSetXLogInsertCheckRecovery();
+
+ if (MyBackendType == B_WAL_SENDER)
+ am_cascading_walsender = true;
+}
+
/*
* IA64-specific code to fetch the AR.BSP register for stack depth checks.
@@ -4224,6 +4261,12 @@ PostgresMain(int argc, char *argv[],
/* Send out notify signals and transmit self-notifies */
ProcessCompletedNotifies();
+ if (DemotePending) {
+ LocalSetXLogInsertNotAllowed();
+ DemotePending = false;
+ SendPostmasterSignal(PMSIGNAL_ADVANCE_STATE_MACHINE);
+ }
+
/*
* Also process incoming notifies, if any. This is mostly to
* ensure stable behavior in tests: if any notifies were
@@ -4285,6 +4328,7 @@ PostgresMain(int argc, char *argv[],
{
ConfigReloadPending = false;
ProcessConfigFile(PGC_SIGHUP);
+ SetLocalRecoveryInProgress();
}
/*
diff --git a/src/backend/utils/init/globals.c b/src/backend/utils/init/globals.c
index 6ab8216839..021f6af434 100644
--- a/src/backend/utils/init/globals.c
+++ b/src/backend/utils/init/globals.c
@@ -33,6 +33,7 @@ volatile sig_atomic_t ProcDiePending = false;
volatile sig_atomic_t ClientConnectionLost = false;
volatile sig_atomic_t IdleInTransactionSessionTimeoutPending = false;
volatile sig_atomic_t ProcSignalBarrierPending = false;
+volatile sig_atomic_t DemotePending = false;
volatile uint32 InterruptHoldoffCount = 0;
volatile uint32 QueryCancelHoldoffCount = 0;
volatile uint32 CritSectionCount = 0;
diff --git a/src/bin/pg_controldata/pg_controldata.c b/src/bin/pg_controldata/pg_controldata.c
index 3e00ac0f70..9ef133f79a 100644
--- a/src/bin/pg_controldata/pg_controldata.c
+++ b/src/bin/pg_controldata/pg_controldata.c
@@ -57,6 +57,8 @@ dbState(DBState state)
return _("shut down");
case DB_SHUTDOWNED_IN_RECOVERY:
return _("shut down in recovery");
+ case DB_DEMOTING:
+ return _("demoting");
case DB_SHUTDOWNING:
return _("shutting down");
case DB_IN_CRASH_RECOVERY:
diff --git a/src/bin/pg_ctl/pg_ctl.c b/src/bin/pg_ctl/pg_ctl.c
index 1cdc3ebaa3..a7805bd219 100644
--- a/src/bin/pg_ctl/pg_ctl.c
+++ b/src/bin/pg_ctl/pg_ctl.c
@@ -62,6 +62,7 @@ typedef enum
RESTART_COMMAND,
RELOAD_COMMAND,
STATUS_COMMAND,
+ DEMOTE_COMMAND,
PROMOTE_COMMAND,
LOGROTATE_COMMAND,
KILL_COMMAND,
@@ -103,6 +104,7 @@ static char version_file[MAXPGPATH];
static char pid_file[MAXPGPATH];
static char backup_file[MAXPGPATH];
static char promote_file[MAXPGPATH];
+static char demote_file[MAXPGPATH];
static char logrotate_file[MAXPGPATH];
static volatile pgpid_t postmasterPID = -1;
@@ -129,6 +131,7 @@ static void do_stop(void);
static void do_restart(void);
static void do_reload(void);
static void do_status(void);
+static void do_demote(void);
static void do_promote(void);
static void do_logrotate(void);
static void do_kill(pgpid_t pid);
@@ -1029,6 +1032,115 @@ do_stop(void)
}
+static void
+do_demote(void)
+{
+ int cnt;
+ FILE *dmtfile;
+ pgpid_t pid;
+ struct stat statbuf;
+
+ pid = get_pgpid(false);
+
+ if (pid == 0) /* no pid file */
+ {
+ write_stderr(_("%s: PID file \"%s\" does not exist\n"), progname, pid_file);
+ write_stderr(_("Is server running?\n"));
+ exit(1);
+ }
+ else if (pid < 0) /* standalone backend, not postmaster */
+ {
+ pid = -pid;
+ write_stderr(_("%s: cannot demote server; "
+ "single-user server is running (PID: %ld)\n"),
+ progname, pid);
+ exit(1);
+ }
+
+ if (shutdown_mode == IMMEDIATE_MODE)
+ {
+ write_stderr(_("%s: cannot demote server using immediate mode"),
+ progname);
+ exit(1);
+ }
+ else if (shutdown_mode == FAST_MODE)
+ snprintf(demote_file, MAXPGPATH, "%s/demote_fast", pg_data);
+ else
+ snprintf(demote_file, MAXPGPATH, "%s/demote", pg_data);
+
+ if ((dmtfile = fopen(demote_file, "w")) == NULL)
+ {
+ write_stderr(_("%s: could not create demote signal file \"%s\": %s\n"),
+ progname, demote_file, strerror(errno));
+ exit(1);
+ }
+
+ if (fclose(dmtfile))
+ {
+ write_stderr(_("%s: could not write demote signal file \"%s\": %s\n"),
+ progname, demote_file, strerror(errno));
+ exit(1);
+ }
+
+ sig = SIGUSR1;
+ if (kill((pid_t) pid, sig) != 0)
+ {
+ write_stderr(_("%s: could not send demote signal (PID: %ld): %s\n"), progname, pid,
+ strerror(errno));
+ exit(1);
+ }
+
+ if (!do_wait)
+ {
+ print_msg(_("server demoting\n"));
+ return;
+ }
+ else
+ {
+ /*
+ * FIXME demote
+ * If backup_label exists, an online backup is running. Warn the user
+ * that smart demote will wait for it to finish. However, if the
+ * server is in archive recovery, we're recovering from an online
+ * backup instead of performing one.
+ */
+ if (shutdown_mode == SMART_MODE &&
+ stat(backup_file, &statbuf) == 0 &&
+ get_control_dbstate() != DB_IN_ARCHIVE_RECOVERY)
+ {
+ print_msg(_("WARNING: online backup mode is active\n"
+ "Demote will not complete until pg_stop_backup() is called.\n\n"));
+ }
+
+ print_msg(_("waiting for server to demote..."));
+
+ for (cnt = 0; cnt < wait_seconds * WAITS_PER_SEC; cnt++)
+ {
+ if (get_control_dbstate() == DB_IN_ARCHIVE_RECOVERY)
+ break;
+
+ if (cnt % WAITS_PER_SEC == 0)
+ print_msg(".");
+ pg_usleep(USEC_PER_SEC / WAITS_PER_SEC);
+ }
+
+ if (get_control_dbstate() != DB_IN_ARCHIVE_RECOVERY)
+ {
+ print_msg(_(" failed\n"));
+
+ write_stderr(_("%s: server does not demote\n"), progname);
+ if (shutdown_mode == SMART_MODE)
+ write_stderr(_("HINT: The \"-m fast\" option immediately disconnects sessions rather than\n"
+ "waiting for session-initiated disconnection.\n"));
+ exit(1);
+ }
+ print_msg(_(" done\n"));
+
+ print_msg(_("server demoted\n"));
+ }
+}
+
+
/*
* restart/reload routines
*/
@@ -2447,6 +2559,8 @@ main(int argc, char **argv)
ctl_command = RELOAD_COMMAND;
else if (strcmp(argv[optind], "status") == 0)
ctl_command = STATUS_COMMAND;
+ else if (strcmp(argv[optind], "demote") == 0)
+ ctl_command = DEMOTE_COMMAND;
else if (strcmp(argv[optind], "promote") == 0)
ctl_command = PROMOTE_COMMAND;
else if (strcmp(argv[optind], "logrotate") == 0)
@@ -2554,6 +2668,9 @@ main(int argc, char **argv)
case RELOAD_COMMAND:
do_reload();
break;
+ case DEMOTE_COMMAND:
+ do_demote();
+ break;
case PROMOTE_COMMAND:
do_promote();
break;
diff --git a/src/include/access/twophase.h b/src/include/access/twophase.h
index 2ca71c3445..4b56f92181 100644
--- a/src/include/access/twophase.h
+++ b/src/include/access/twophase.h
@@ -53,6 +53,7 @@ extern void RecoverPreparedTransactions(void);
extern void CheckPointTwoPhase(XLogRecPtr redo_horizon);
extern void FinishPreparedTransaction(const char *gid, bool isCommit);
+void ShutdownPreparedTransactions(void);
extern void PrepareRedoAdd(char *buf, XLogRecPtr start_lsn,
XLogRecPtr end_lsn, RepOriginId origin_id);
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 8c9cadc6da..b1b1ea67f9 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -219,18 +219,20 @@ extern bool XLOG_DEBUG;
/* These directly affect the behavior of CreateCheckPoint and subsidiaries */
#define CHECKPOINT_IS_SHUTDOWN 0x0001 /* Checkpoint is for shutdown */
-#define CHECKPOINT_END_OF_RECOVERY 0x0002 /* Like shutdown checkpoint, but
+#define CHECKPOINT_IS_DEMOTE 0x0002 /* Like shutdown checkpoint, but
+ * issued at end of WAL production */
+#define CHECKPOINT_END_OF_RECOVERY 0x0004 /* Like shutdown checkpoint, but
* issued at end of WAL recovery */
-#define CHECKPOINT_IMMEDIATE 0x0004 /* Do it without delays */
-#define CHECKPOINT_FORCE 0x0008 /* Force even if no activity */
-#define CHECKPOINT_FLUSH_ALL 0x0010 /* Flush all pages, including those
+#define CHECKPOINT_IMMEDIATE 0x0008 /* Do it without delays */
+#define CHECKPOINT_FORCE 0x0010 /* Force even if no activity */
+#define CHECKPOINT_FLUSH_ALL 0x0020 /* Flush all pages, including those
* belonging to unlogged tables */
/* These are important to RequestCheckpoint */
-#define CHECKPOINT_WAIT 0x0020 /* Wait for completion */
-#define CHECKPOINT_REQUESTED 0x0040 /* Checkpoint request has been made */
+#define CHECKPOINT_WAIT 0x0040 /* Wait for completion */
+#define CHECKPOINT_REQUESTED 0x0080 /* Checkpoint request has been made */
/* These indicate the cause of a checkpoint request */
-#define CHECKPOINT_CAUSE_XLOG 0x0080 /* XLOG consumption */
-#define CHECKPOINT_CAUSE_TIME 0x0100 /* Elapsed time */
+#define CHECKPOINT_CAUSE_XLOG 0x0100 /* XLOG consumption */
+#define CHECKPOINT_CAUSE_TIME 0x0200 /* Elapsed time */
/*
* Flag bits for the record being inserted, set using XLogSetRecordFlags().
@@ -301,6 +303,7 @@ extern const char *xlog_identify(uint8 info);
extern void issue_xlog_fsync(int fd, XLogSegNo segno);
+extern bool SetLocalRecoveryInProgress(void);
extern bool RecoveryInProgress(void);
extern RecoveryState GetRecoveryState(void);
extern bool HotStandbyActive(void);
@@ -397,4 +400,8 @@ extern SessionBackupState get_backup_status(void);
/* files to signal promotion to primary */
#define PROMOTE_SIGNAL_FILE "promote"
+/* files to signal demotion to standby */
+#define DEMOTE_SIGNAL_FILE "demote"
+#define DEMOTE_FAST_SIGNAL_FILE "demote_fast"
+
#endif /* XLOG_H */
diff --git a/src/include/catalog/pg_control.h b/src/include/catalog/pg_control.h
index 06bed90c5e..3b30c1d767 100644
--- a/src/include/catalog/pg_control.h
+++ b/src/include/catalog/pg_control.h
@@ -87,6 +87,7 @@ typedef enum DBState
DB_STARTUP = 0,
DB_SHUTDOWNED,
DB_SHUTDOWNED_IN_RECOVERY,
+ DB_DEMOTING,
DB_SHUTDOWNING,
DB_IN_CRASH_RECOVERY,
DB_IN_ARCHIVE_RECOVERY,
diff --git a/src/include/libpq/libpq-be.h b/src/include/libpq/libpq-be.h
index 0a23281ad5..24ba3da013 100644
--- a/src/include/libpq/libpq-be.h
+++ b/src/include/libpq/libpq-be.h
@@ -70,7 +70,7 @@ typedef struct
typedef enum CAC_state
{
- CAC_OK, CAC_STARTUP, CAC_SHUTDOWN, CAC_RECOVERY, CAC_TOOMANY,
+ CAC_OK, CAC_STARTUP, CAC_DEMOTE, CAC_SHUTDOWN, CAC_RECOVERY, CAC_TOOMANY,
CAC_SUPERUSER
} CAC_state;
diff --git a/src/include/miscadmin.h b/src/include/miscadmin.h
index 72e3352398..d60804208f 100644
--- a/src/include/miscadmin.h
+++ b/src/include/miscadmin.h
@@ -83,6 +83,7 @@ extern PGDLLIMPORT volatile sig_atomic_t QueryCancelPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcDiePending;
extern PGDLLIMPORT volatile sig_atomic_t IdleInTransactionSessionTimeoutPending;
extern PGDLLIMPORT volatile sig_atomic_t ProcSignalBarrierPending;
+extern PGDLLIMPORT volatile sig_atomic_t DemotePending;
extern PGDLLIMPORT volatile sig_atomic_t ClientConnectionLost;
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 1387201382..f1c0a37e76 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -880,6 +880,7 @@ typedef enum
WAIT_EVENT_PROCARRAY_GROUP_UPDATE,
WAIT_EVENT_PROC_SIGNAL_BARRIER,
WAIT_EVENT_PROMOTE,
+ WAIT_EVENT_DEMOTE,
WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT,
WAIT_EVENT_RECOVERY_CONFLICT_TABLESPACE,
WAIT_EVENT_RECOVERY_PAUSE,
diff --git a/src/include/postmaster/bgwriter.h b/src/include/postmaster/bgwriter.h
index 0a5708b32e..4d4f0ea1dd 100644
--- a/src/include/postmaster/bgwriter.h
+++ b/src/include/postmaster/bgwriter.h
@@ -41,5 +41,6 @@ extern Size CheckpointerShmemSize(void);
extern void CheckpointerShmemInit(void);
extern bool FirstCallSinceLastCheckpoint(void);
+extern void ReqCheckpointDemoteHandler(SIGNAL_ARGS);
#endif /* _BGWRITER_H */
diff --git a/src/include/storage/lock.h b/src/include/storage/lock.h
index 1c3e9c1999..fa8d64e1a7 100644
--- a/src/include/storage/lock.h
+++ b/src/include/storage/lock.h
@@ -584,6 +584,8 @@ extern void lock_twophase_postcommit(TransactionId xid, uint16 info,
void *recdata, uint32 len);
extern void lock_twophase_postabort(TransactionId xid, uint16 info,
void *recdata, uint32 len);
+extern void lock_twophase_shutdown(TransactionId xid, uint16 info,
+ void *recdata, uint32 len);
extern void lock_twophase_standby_recover(TransactionId xid, uint16 info,
void *recdata, uint32 len);
diff --git a/src/include/storage/procsignal.h b/src/include/storage/procsignal.h
index 5cb39697f3..7264e9a705 100644
--- a/src/include/storage/procsignal.h
+++ b/src/include/storage/procsignal.h
@@ -34,6 +34,10 @@ typedef enum
PROCSIG_PARALLEL_MESSAGE, /* message from cooperating parallel backend */
PROCSIG_WALSND_INIT_STOPPING, /* ask walsenders to prepare for shutdown */
PROCSIG_BARRIER, /* global barrier interrupt */
+ PROCSIG_DEMOTING, /* ask backends to demote in smart mode */
+ PROCSIG_DEMOTING_FAST, /* ask backends to demote in fast mode */
+ PROCSIG_DEMOTED, /* ask backends to switch to recovery mode */
+ PROCSIG_CHECKPOINTER_DEMOTING, /* ask checkpointer to demote */
/* Recovery conflict reasons */
PROCSIG_RECOVERY_CONFLICT_DATABASE,
diff --git a/src/include/tcop/tcopprot.h b/src/include/tcop/tcopprot.h
index bd30607b07..e5f42f9fec 100644
--- a/src/include/tcop/tcopprot.h
+++ b/src/include/tcop/tcopprot.h
@@ -68,6 +68,8 @@ extern void StatementCancelHandler(SIGNAL_ARGS);
extern void FloatExceptionHandler(SIGNAL_ARGS) pg_attribute_noreturn();
extern void RecoveryConflictInterrupt(ProcSignalReason reason); /* called from SIGUSR1
* handler */
+extern void ReqDemoteHandler(ProcSignalReason reason); /* called from SIGUSR1 handler */
+extern void ReqDemotedHandler(ProcSignalReason reason); /* called from SIGUSR1 handler */
extern void ProcessClientReadInterrupt(bool blocked);
extern void ProcessClientWriteInterrupt(bool blocked);
diff --git a/src/include/utils/pidfile.h b/src/include/utils/pidfile.h
index 63fefe5c4c..f761d2c4ef 100644
--- a/src/include/utils/pidfile.h
+++ b/src/include/utils/pidfile.h
@@ -50,6 +50,7 @@
*/
#define PM_STATUS_STARTING "starting" /* still starting up */
#define PM_STATUS_STOPPING "stopping" /* in shutdown sequence */
+#define PM_STATUS_DEMOTING "demoting" /* demote sequence */
#define PM_STATUS_READY "ready " /* ready for connections */
#define PM_STATUS_STANDBY "standby " /* up, won't accept connections */
--
2.20.1
--MP_/6xG3U2Dzv0UR=.BCNK7fiL6
Content-Type: text/x-patch
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename=v5-0003-demote-add-pg_demote-function.patch
^ permalink raw reply [nested|flat] 2+ messages in thread
* [PATCH v53 1/7] Introduce an option to make logical replication database specific.
@ 2026-04-04 15:16 Antonin Houska <[email protected]>
0 siblings, 0 replies; 2+ messages in thread
From: Antonin Houska @ 2026-04-04 15:16 UTC (permalink / raw)
By default, the logical decoding assumes access to shared catalogs, so the
snapshot builder needs to consider cluster-wide XIDs during startup. That in
turn means that, if any transaction is already running (and has XID assigned),
the snapshot builder needs to wait for its completion, as it does not know if
that transaction performed catalog changes earlier.
Possible problem with this concept is that if REPACK (CONCURRENTLY) is running
in some database, backends running the same command in other databases get
stuck until the first one has committed. Thus only as single backend in the
cluster can run REPACK (CONCURRENTLY) at any time. Likewise, REPACK
(CONCURRENTLY) can block walsenders starting on behalf of subscriptions
throughout the cluster.
This patch adds a new option to logical replication output plugin, to declare
that it does not use shared catalogs (i.e. catalogs that can be changed by
transactions running in other databases in the cluster). In that case, no
snapshot the backend will use during the decoding needs to contain information
about transactions running in other databases. Thus the snapshot builder only
needs to wait for completion of transactions in the current database.
Currently we only use this option in the REPACK background worker. It could
possibly be used in the plugin for logical replication too, however that would
need thorough analysis of that plugin.
The patch bumps WAL version number, due to a new field in xl_running_xacts.
---
contrib/pg_visibility/pg_visibility.c | 4 +-
doc/src/sgml/logicaldecoding.sgml | 4 ++
src/backend/access/index/genam.c | 8 +++
src/backend/access/rmgrdesc/standbydesc.c | 2 +
src/backend/access/transam/xlog.c | 2 +-
src/backend/access/transam/xlogfuncs.c | 2 +-
src/backend/postmaster/bgwriter.c | 2 +-
src/backend/replication/logical/decode.c | 8 ++-
src/backend/replication/logical/logical.c | 3 ++
src/backend/replication/logical/snapbuild.c | 59 ++++++++++++++++++---
src/backend/replication/slot.c | 2 +-
src/backend/storage/ipc/procarray.c | 23 +++++++-
src/backend/storage/ipc/standby.c | 14 ++++-
src/include/access/genam.h | 7 +++
src/include/access/xlog_internal.h | 2 +-
src/include/replication/output_plugin.h | 1 +
src/include/replication/snapbuild.h | 3 +-
src/include/storage/procarray.h | 2 +-
src/include/storage/standby.h | 3 +-
src/include/storage/standbydefs.h | 1 +
20 files changed, 131 insertions(+), 21 deletions(-)
diff --git a/contrib/pg_visibility/pg_visibility.c b/contrib/pg_visibility/pg_visibility.c
index dfab0b64cf5..d564bd2a00c 100644
--- a/contrib/pg_visibility/pg_visibility.c
+++ b/contrib/pg_visibility/pg_visibility.c
@@ -621,7 +621,7 @@ GetStrictOldestNonRemovableTransactionId(Relation rel)
else if (rel == NULL || rel->rd_rel->relisshared)
{
/* Shared relation: take into account all running xids */
- runningTransactions = GetRunningTransactionData();
+ runningTransactions = GetRunningTransactionData(InvalidOid);
LWLockRelease(ProcArrayLock);
LWLockRelease(XidGenLock);
return runningTransactions->oldestRunningXid;
@@ -632,7 +632,7 @@ GetStrictOldestNonRemovableTransactionId(Relation rel)
* Normal relation: take into account xids running within the current
* database
*/
- runningTransactions = GetRunningTransactionData();
+ runningTransactions = GetRunningTransactionData(InvalidOid);
LWLockRelease(ProcArrayLock);
LWLockRelease(XidGenLock);
return runningTransactions->oldestDatabaseRunningXid;
diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index 6dc49108997..28089d1053c 100644
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -959,6 +959,7 @@ typedef struct OutputPluginOptions
{
OutputPluginOutputType output_type;
bool receive_rewrites;
+ bool need_shared_catalogs;
} OutputPluginOptions;
</programlisting>
<literal>output_type</literal> has to either be set to
@@ -969,6 +970,9 @@ typedef struct OutputPluginOptions
also be called for changes made by heap rewrites during certain DDL
operations. These are of interest to plugins that handle DDL
replication, but they require special handling.
+ <literal>need_shared_catalogs</literal> can be set to false if you are
+ sure the plugin functions do not access shared system catalogs. It can
+ speed up creation of replication slots that use this plugin.
</para>
<para>
diff --git a/src/backend/access/index/genam.c b/src/backend/access/index/genam.c
index 1408989c568..44df2605068 100644
--- a/src/backend/access/index/genam.c
+++ b/src/backend/access/index/genam.c
@@ -394,6 +394,14 @@ systable_beginscan(Relation heapRelation,
SysScanDesc sysscan;
Relation irel;
+ /*
+ * If this backend promised that it won't access shared catalogs during
+ * logical decoding, this seems to be the right place to check.
+ */
+ Assert(!HistoricSnapshotActive() ||
+ accessSharedCatalogsInDecoding ||
+ !heapRelation->rd_rel->relisshared);
+
if (indexOK &&
!IgnoreSystemIndexes &&
!ReindexIsProcessingIndex(indexId))
diff --git a/src/backend/access/rmgrdesc/standbydesc.c b/src/backend/access/rmgrdesc/standbydesc.c
index 0a291354ae2..685d1bdb024 100644
--- a/src/backend/access/rmgrdesc/standbydesc.c
+++ b/src/backend/access/rmgrdesc/standbydesc.c
@@ -41,6 +41,8 @@ standby_desc_running_xacts(StringInfo buf, xl_running_xacts *xlrec)
for (i = 0; i < xlrec->subxcnt; i++)
appendStringInfo(buf, " %u", xlrec->xids[xlrec->xcnt + i]);
}
+
+ appendStringInfo(buf, "; dbid: %u", xlrec->dbid);
}
void
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 9e8999bbb61..1a6c1456d2c 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -7731,7 +7731,7 @@ CreateCheckPoint(int flags)
* recovery we don't need to write running xact data.
*/
if (!shutdown && XLogStandbyInfoActive())
- LogStandbySnapshot();
+ LogStandbySnapshot(InvalidOid);
START_CRIT_SECTION();
diff --git a/src/backend/access/transam/xlogfuncs.c b/src/backend/access/transam/xlogfuncs.c
index 65bbaeda59c..0f5979691e6 100644
--- a/src/backend/access/transam/xlogfuncs.c
+++ b/src/backend/access/transam/xlogfuncs.c
@@ -245,7 +245,7 @@ pg_log_standby_snapshot(PG_FUNCTION_ARGS)
(errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE),
errmsg("pg_log_standby_snapshot() can only be used if \"wal_level\" >= \"replica\"")));
- recptr = LogStandbySnapshot();
+ recptr = LogStandbySnapshot(InvalidOid);
/*
* As a convenience, return the WAL location of the last inserted record
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 1d8947774a9..a30de4262eb 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -289,7 +289,7 @@ BackgroundWriterMain(const void *startup_data, size_t startup_data_len)
if (now >= timeout &&
last_snapshot_lsn <= GetLastImportantRecPtr())
{
- last_snapshot_lsn = LogStandbySnapshot();
+ last_snapshot_lsn = LogStandbySnapshot(InvalidOid);
last_snapshot_ts = now;
}
}
diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c
index 57aaef57c61..4299d0bc867 100644
--- a/src/backend/replication/logical/decode.c
+++ b/src/backend/replication/logical/decode.c
@@ -381,7 +381,13 @@ standby_decode(LogicalDecodingContext *ctx, XLogRecordBuffer *buf)
{
xl_running_xacts *running = (xl_running_xacts *) XLogRecGetData(r);
- SnapBuildProcessRunningXacts(builder, buf->origptr, running);
+
+ /*
+ * If the plugin does not access shared catalogs, only keep
+ * track of transactions in the current database.
+ */
+ SnapBuildProcessRunningXacts(builder, buf->origptr, running,
+ !ctx->options.need_shared_catalogs);
/*
* Abort all transactions that we keep track of, that are
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index f20d0c542f3..8ceaf64d164 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -285,6 +285,9 @@ StartupDecodingContext(List *output_plugin_options,
ctx->write = do_write;
ctx->update_progress = update_progress;
+ /* Assume shared catalog access. The startup callback can change it. */
+ ctx->options.need_shared_catalogs = true;
+
ctx->output_plugin_options = output_plugin_options;
ctx->fast_forward = fast_forward;
diff --git a/src/backend/replication/logical/snapbuild.c b/src/backend/replication/logical/snapbuild.c
index b4269a3b102..a6738e120c7 100644
--- a/src/backend/replication/logical/snapbuild.c
+++ b/src/backend/replication/logical/snapbuild.c
@@ -125,6 +125,7 @@
#include <sys/stat.h>
#include <unistd.h>
+#include "access/genam.h"
#include "access/heapam_xlog.h"
#include "access/transam.h"
#include "access/xact.h"
@@ -154,6 +155,14 @@
static ResourceOwner SavedResourceOwnerDuringExport = NULL;
static bool ExportInProgress = false;
+/*
+ * If a backend is going to do logical decoding and the output plugin does
+ * not need to access shared catalogs, setting this variable to false can make
+ * the decoding startup faster. In particular, the backend will not need to
+ * wait for completion of already running transactions in other databases.
+ */
+bool accessSharedCatalogsInDecoding = true;
+
/* ->committed and ->catchange manipulation */
static void SnapBuildPurgeOlderTxn(SnapBuild *builder);
@@ -170,7 +179,8 @@ static inline bool SnapBuildXidHasCatalogChanges(SnapBuild *builder, Transaction
uint32 xinfo);
/* xlog reading helper functions for SnapBuildProcessRunningXacts */
-static bool SnapBuildFindSnapshot(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *running);
+static bool SnapBuildFindSnapshot(SnapBuild *builder, XLogRecPtr lsn,
+ xl_running_xacts *running, bool db_specific);
static void SnapBuildWaitSnapshot(xl_running_xacts *running, TransactionId cutoff);
/* serialization functions */
@@ -1136,7 +1146,8 @@ SnapBuildXidHasCatalogChanges(SnapBuild *builder, TransactionId xid,
* anymore.
*/
void
-SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *running)
+SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *running,
+ bool db_specific)
{
ReorderBufferTXN *txn;
TransactionId xmin;
@@ -1149,7 +1160,7 @@ SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xact
if (builder->state < SNAPBUILD_CONSISTENT)
{
/* returns false if there's no point in performing cleanup just yet */
- if (!SnapBuildFindSnapshot(builder, lsn, running))
+ if (!SnapBuildFindSnapshot(builder, lsn, running, db_specific))
return;
}
else
@@ -1166,8 +1177,14 @@ SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xact
* (see SnapBuildCommitTxn). Because of this, xmax can be lower than
* xmin, which looks odd but is correct and actually more efficient, since
* we hit fast paths in heapam_visibility.c.
+ *
+ * If database specific transaction info was used during startup, the info
+ * for the whole cluster can make xmin go backwards. That would be bad
+ * because we might no longer have older XIDs in ->committed.
*/
- builder->xmin = running->oldestRunningXid;
+ if (!db_specific ||
+ NormalTransactionIdFollows(running->oldestRunningXid, builder->xmin))
+ builder->xmin = running->oldestRunningXid;
/* Remove transactions we don't need to keep track off anymore */
SnapBuildPurgeOlderTxn(builder);
@@ -1182,7 +1199,11 @@ SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xact
* the record we're reading now.
*/
xmin = ReorderBufferGetOldestXmin(builder->reorder);
- if (xmin == InvalidTransactionId)
+
+ /*
+ * Like above, do not let slot xmin go backwards.
+ */
+ if (xmin == InvalidTransactionId && !db_specific)
xmin = running->oldestRunningXid;
elog(DEBUG3, "xmin: %u, xmax: %u, oldest running: %u, oldest xmin: %u",
builder->xmin, builder->xmax, running->oldestRunningXid, xmin);
@@ -1238,7 +1259,8 @@ SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn, xl_running_xact
* using the xl_running_xacts record.
*/
static bool
-SnapBuildFindSnapshot(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *running)
+SnapBuildFindSnapshot(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *running,
+ bool db_specific)
{
/* ---
* Build catalog decoding snapshot incrementally using information about
@@ -1265,6 +1287,25 @@ SnapBuildFindSnapshot(SnapBuild *builder, XLogRecPtr lsn, xl_running_xacts *runn
* ---
*/
+ if (db_specific)
+ {
+ /*
+ * If we must only keep track of transactions running in the current
+ * database, we need transaction info from exactly that database.
+ */
+ if (running->dbid != MyDatabaseId)
+ {
+ LogStandbySnapshot(MyDatabaseId);
+ return false;
+ }
+
+ /*
+ * We'd better be able to check during scan if the plugin does not
+ * lie.
+ */
+ accessSharedCatalogsInDecoding = false;
+ }
+
/*
* xl_running_xacts record is older than what we can use, we might not
* have all necessary catalog rows anymore.
@@ -1465,7 +1506,11 @@ SnapBuildWaitSnapshot(xl_running_xacts *running, TransactionId cutoff)
*/
if (!RecoveryInProgress())
{
- LogStandbySnapshot();
+ /*
+ * If the last transaction info was about specific database, so needs
+ * to be the next one - at least until we're in the CONSISTENT state.
+ */
+ LogStandbySnapshot(running->dbid);
}
}
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index a9092fc2382..9533515a63b 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -1762,7 +1762,7 @@ ReplicationSlotReserveWal(void)
XLogRecPtr flushptr;
/* make sure we have enough information to start */
- flushptr = LogStandbySnapshot();
+ flushptr = LogStandbySnapshot(InvalidOid);
/* and make sure it's fsynced to disk */
XLogFlush(flushptr);
diff --git a/src/backend/storage/ipc/procarray.c b/src/backend/storage/ipc/procarray.c
index cc207cb56e3..92ed34128e5 100644
--- a/src/backend/storage/ipc/procarray.c
+++ b/src/backend/storage/ipc/procarray.c
@@ -2631,9 +2631,11 @@ ProcArrayInstallRestoredXmin(TransactionId xmin, PGPROC *proc)
*
* Note that if any transaction has overflowed its cached subtransactions
* then there is no real need include any subtransactions.
+ *
+ * If 'dbid' is valid, only gather transactions running in that database.
*/
RunningTransactions
-GetRunningTransactionData(void)
+GetRunningTransactionData(Oid dbid)
{
/* result workspace */
static RunningTransactionsData CurrentRunningXactsData;
@@ -2708,6 +2710,18 @@ GetRunningTransactionData(void)
if (!TransactionIdIsValid(xid))
continue;
+ /*
+ * Filter by database OID if requested.
+ */
+ if (OidIsValid(dbid))
+ {
+ int pgprocno = arrayP->pgprocnos[index];
+ PGPROC *proc = &allProcs[pgprocno];
+
+ if (proc->databaseId != dbid)
+ continue;
+ }
+
/*
* Be careful not to exclude any xids before calculating the values of
* oldestRunningXid and suboverflowed, since these are used to clean
@@ -2758,6 +2772,12 @@ GetRunningTransactionData(void)
PGPROC *proc = &allProcs[pgprocno];
int nsubxids;
+ /*
+ * Filter by database OID if requested.
+ */
+ if (OidIsValid(dbid) && proc->databaseId != dbid)
+ continue;
+
/*
* Save subtransaction XIDs. Other backends can't add or remove
* entries while we're holding XidGenLock.
@@ -2791,6 +2811,7 @@ GetRunningTransactionData(void)
* increases if slots do.
*/
+ CurrentRunningXacts->dbid = dbid;
CurrentRunningXacts->xcnt = count - subcount;
CurrentRunningXacts->subxcnt = subcount;
CurrentRunningXacts->subxid_status = suboverflowed ? SUBXIDS_IN_SUBTRANS : SUBXIDS_IN_ARRAY;
diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
index de9092fdf5b..c653ea742bc 100644
--- a/src/backend/storage/ipc/standby.c
+++ b/src/backend/storage/ipc/standby.c
@@ -1188,6 +1188,14 @@ standby_redo(XLogReaderState *record)
xl_running_xacts *xlrec = (xl_running_xacts *) XLogRecGetData(record);
RunningTransactionsData running;
+ /*
+ * Records issued for specific database are not suitable for physical
+ * replication because that affects the whole cluster. In particular,
+ * the list of XID is probably incomplete here.
+ */
+ if (OidIsValid(xlrec->dbid))
+ return;
+
running.xcnt = xlrec->xcnt;
running.subxcnt = xlrec->subxcnt;
running.subxid_status = xlrec->subxid_overflow ? SUBXIDS_MISSING : SUBXIDS_IN_ARRAY;
@@ -1277,11 +1285,12 @@ standby_redo(XLogReaderState *record)
* as there's no independent knob to just enable logical decoding. For
* details of how this is used, check snapbuild.c's introductory comment.
*
+ * If 'dbid' is valid, only gather transactions running in that database.
*
* Returns the RecPtr of the last inserted record.
*/
XLogRecPtr
-LogStandbySnapshot(void)
+LogStandbySnapshot(Oid dbid)
{
XLogRecPtr recptr;
RunningTransactions running;
@@ -1314,7 +1323,7 @@ LogStandbySnapshot(void)
* Log details of all in-progress transactions. This should be the last
* record we write, because standby will open up when it sees this.
*/
- running = GetRunningTransactionData();
+ running = GetRunningTransactionData(dbid);
/*
* GetRunningTransactionData() acquired ProcArrayLock, we must release it.
@@ -1358,6 +1367,7 @@ LogCurrentRunningXacts(RunningTransactions CurrRunningXacts)
xl_running_xacts xlrec;
XLogRecPtr recptr;
+ xlrec.dbid = CurrRunningXacts->dbid;
xlrec.xcnt = CurrRunningXacts->xcnt;
xlrec.subxcnt = CurrRunningXacts->subxcnt;
xlrec.subxid_overflow = (CurrRunningXacts->subxid_status != SUBXIDS_IN_ARRAY);
diff --git a/src/include/access/genam.h b/src/include/access/genam.h
index b69320a7fc8..c9a52277be1 100644
--- a/src/include/access/genam.h
+++ b/src/include/access/genam.h
@@ -136,6 +136,13 @@ typedef struct IndexOrderByDistance
bool isnull;
} IndexOrderByDistance;
+
+/*
+ * Keep track of whether logical decoding in this backend promised not to
+ * access shared catalogs, as a safety check. Somewhat misplaced, but ...
+ */
+extern bool accessSharedCatalogsInDecoding;
+
/*
* generalized index_ interface routines (in indexam.c)
*/
diff --git a/src/include/access/xlog_internal.h b/src/include/access/xlog_internal.h
index 10c18d39ff8..13ae3ad4fbb 100644
--- a/src/include/access/xlog_internal.h
+++ b/src/include/access/xlog_internal.h
@@ -32,7 +32,7 @@
/*
* Each page of XLOG file has a header like this:
*/
-#define XLOG_PAGE_MAGIC 0xD11E /* can be used as WAL version indicator */
+#define XLOG_PAGE_MAGIC 0xD11F /* can be used as WAL version indicator */
typedef struct XLogPageHeaderData
{
diff --git a/src/include/replication/output_plugin.h b/src/include/replication/output_plugin.h
index 842fcde67f9..917f3cff232 100644
--- a/src/include/replication/output_plugin.h
+++ b/src/include/replication/output_plugin.h
@@ -27,6 +27,7 @@ typedef struct OutputPluginOptions
{
OutputPluginOutputType output_type;
bool receive_rewrites;
+ bool need_shared_catalogs;
} OutputPluginOptions;
/*
diff --git a/src/include/replication/snapbuild.h b/src/include/replication/snapbuild.h
index a22a83a2f23..d02530a912a 100644
--- a/src/include/replication/snapbuild.h
+++ b/src/include/replication/snapbuild.h
@@ -92,7 +92,8 @@ extern void SnapBuildProcessNewCid(SnapBuild *builder, TransactionId xid,
XLogRecPtr lsn,
xl_heap_new_cid *xlrec);
extern void SnapBuildProcessRunningXacts(SnapBuild *builder, XLogRecPtr lsn,
- xl_running_xacts *running);
+ xl_running_xacts *running,
+ bool db_specific);
extern void SnapBuildSerializationPoint(SnapBuild *builder, XLogRecPtr lsn);
extern bool SnapBuildSnapshotExists(XLogRecPtr lsn);
diff --git a/src/include/storage/procarray.h b/src/include/storage/procarray.h
index abdf021e66e..377b3060b9f 100644
--- a/src/include/storage/procarray.h
+++ b/src/include/storage/procarray.h
@@ -49,7 +49,7 @@ extern bool ProcArrayInstallImportedXmin(TransactionId xmin,
VirtualTransactionId *sourcevxid);
extern bool ProcArrayInstallRestoredXmin(TransactionId xmin, PGPROC *proc);
-extern RunningTransactions GetRunningTransactionData(void);
+extern RunningTransactions GetRunningTransactionData(Oid dbid);
extern bool TransactionIdIsInProgress(TransactionId xid);
extern TransactionId GetOldestNonRemovableTransactionId(Relation rel);
diff --git a/src/include/storage/standby.h b/src/include/storage/standby.h
index 6a314c693cd..8715c08e94f 100644
--- a/src/include/storage/standby.h
+++ b/src/include/storage/standby.h
@@ -126,6 +126,7 @@ typedef enum
typedef struct RunningTransactionsData
{
+ Oid dbid; /* only track xacts in this database */
int xcnt; /* # of xact ids in xids[] */
int subxcnt; /* # of subxact ids in xids[] */
subxids_array_status subxid_status;
@@ -143,7 +144,7 @@ typedef RunningTransactionsData *RunningTransactions;
extern void LogAccessExclusiveLock(Oid dbOid, Oid relOid);
extern void LogAccessExclusiveLockPrepare(void);
-extern XLogRecPtr LogStandbySnapshot(void);
+extern XLogRecPtr LogStandbySnapshot(Oid dbid);
extern void LogStandbyInvalidations(int nmsgs, SharedInvalidationMessage *msgs,
bool relcacheInitFileInval);
diff --git a/src/include/storage/standbydefs.h b/src/include/storage/standbydefs.h
index 231d251fd51..e75b7078766 100644
--- a/src/include/storage/standbydefs.h
+++ b/src/include/storage/standbydefs.h
@@ -46,6 +46,7 @@ typedef struct xl_standby_locks
*/
typedef struct xl_running_xacts
{
+ Oid dbid; /* only track xacts in this database */
int xcnt; /* # of xact ids in xids[] */
int subxcnt; /* # of subxact ids in xids[] */
bool subxid_overflow; /* snapshot overflowed, subxids missing */
--
2.47.3
--qr3jlalmmcpkiodg
Content-Type: text/x-diff; charset=utf-8
Content-Disposition: attachment;
filename="v53-0002-Fix-a-few-problems-in-index-build-progress-repor.patch"
^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2026-04-04 15:16 UTC | newest]
Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2020-04-10 16:01 [PATCH 2/4] demote: support demoting instance from production to standby Jehan-Guillaume de Rorthais <[email protected]>
2026-04-04 15:16 [PATCH v53 1/7] Introduce an option to make logical replication database specific. Antonin Houska <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox