public inbox for [email protected]
help / color / mirror / Atom feed[PATCH 7/7] Move code to apply one WAL record to a subroutine.
18+ messages / 4 participants
[nested] [flat]
* [PATCH 7/7] Move code to apply one WAL record to a subroutine.
@ 2021-06-21 21:00 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-06-21 21:00 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 283 +++++++++++-----------
1 file changed, 148 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 44eb425eaf9..d7787c9a082 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -366,6 +366,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1392,11 +1393,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1418,11 +1416,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
(rmid == RM_XACT_ID && trace_recovery_messages <= DEBUG2) ||
@@ -1432,8 +1425,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1488,132 +1481,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
/*
- * ShmemVariableCache->nextXid must be beyond record's xid.
+ * Apply the record
*/
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newTLI = ThisTimeLineID;
- TimeLineID prevTLI = ThisTimeLineID;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newTLI = checkPoint.ThisTimeLineID;
- prevTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newTLI = xlrec.ThisTimeLineID;
- prevTLI = xlrec.PrevTimeLineID;
- }
-
- if (newTLI != ThisTimeLineID)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
-
- /* Following WAL records should be run with new TLI */
- ThisTimeLineID = newTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->replayEndRecPtr = EndRecPtr;
- XLogRecCtl->replayEndTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
- /*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
- */
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1672,7 +1543,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1701,6 +1572,148 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the
+ * current timeline to change. The record is already considered to
+ * be part of the new timeline, so we update ThisTimeLineID before
+ * replaying it. That's important so that replayEndTLI, which is
+ * recorded as the minimum recovery point's TLI if recovery stops
+ * after this record, is set correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newTLI = ThisTimeLineID;
+ TimeLineID prevTLI = ThisTimeLineID;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newTLI = checkPoint.ThisTimeLineID;
+ prevTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newTLI = xlrec.ThisTimeLineID;
+ prevTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newTLI != ThisTimeLineID)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
+
+ /* Following WAL records should be run with new TLI */
+ ThisTimeLineID = newTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so
+ * that XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecCtl->replayEndTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we
+ * see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the
+ * WAL record are consistent with the existing pages. This check
+ * is done only if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been
+ * successfully replayed.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up
+ * the receiver so that it notices the updated
+ * lastReplayedEndRecPtr and sends a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any
+ * (possibly bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new
+ * timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------4DE063CB5604E65545208F19--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH 7/7] Move code to apply one WAL record to a subroutine.
@ 2021-06-21 21:00 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-06-21 21:00 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 283 +++++++++++-----------
1 file changed, 148 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 5e23244f6da..c78fc5273bd 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -366,6 +366,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1398,11 +1399,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1424,11 +1422,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
(rmid == RM_XACT_ID && trace_recovery_messages <= DEBUG2) ||
@@ -1438,8 +1431,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1494,132 +1487,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
/*
- * ShmemVariableCache->nextXid must be beyond record's xid.
+ * Apply the record
*/
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newTLI = ThisTimeLineID;
- TimeLineID prevTLI = ThisTimeLineID;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newTLI = checkPoint.ThisTimeLineID;
- prevTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newTLI = xlrec.ThisTimeLineID;
- prevTLI = xlrec.PrevTimeLineID;
- }
-
- if (newTLI != ThisTimeLineID)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
-
- /* Following WAL records should be run with new TLI */
- ThisTimeLineID = newTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->replayEndRecPtr = EndRecPtr;
- XLogRecCtl->replayEndTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
- /*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
- */
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1678,7 +1549,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1707,6 +1578,148 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the
+ * current timeline to change. The record is already considered to
+ * be part of the new timeline, so we update ThisTimeLineID before
+ * replaying it. That's important so that replayEndTLI, which is
+ * recorded as the minimum recovery point's TLI if recovery stops
+ * after this record, is set correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newTLI = ThisTimeLineID;
+ TimeLineID prevTLI = ThisTimeLineID;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newTLI = checkPoint.ThisTimeLineID;
+ prevTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newTLI = xlrec.ThisTimeLineID;
+ prevTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newTLI != ThisTimeLineID)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
+
+ /* Following WAL records should be run with new TLI */
+ ThisTimeLineID = newTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so
+ * that XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecCtl->replayEndTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we
+ * see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the
+ * WAL record are consistent with the existing pages. This check
+ * is done only if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been
+ * successfully replayed.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up
+ * the receiver so that it notices the updated
+ * lastReplayedEndRecPtr and sends a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any
+ * (possibly bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new
+ * timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------8C95DA69A2E7D2FC9457CD47--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v4 3/3] Move code to apply one WAL record to a subroutine.
@ 2021-07-31 12:06 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-07-31 12:06 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 283 +++++++++++-----------
1 file changed, 148 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 6030d6fe819..85909c9b686 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -366,6 +366,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1374,11 +1375,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1400,11 +1398,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
(rmid == RM_XACT_ID && trace_recovery_messages <= DEBUG2) ||
@@ -1414,8 +1407,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1470,132 +1463,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
/*
- * ShmemVariableCache->nextXid must be beyond record's xid.
+ * Apply the record
*/
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newTLI = ThisTimeLineID;
- TimeLineID prevTLI = ThisTimeLineID;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newTLI = checkPoint.ThisTimeLineID;
- prevTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newTLI = xlrec.ThisTimeLineID;
- prevTLI = xlrec.PrevTimeLineID;
- }
-
- if (newTLI != ThisTimeLineID)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
-
- /* Following WAL records should be run with new TLI */
- ThisTimeLineID = newTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->replayEndRecPtr = EndRecPtr;
- XLogRecCtl->replayEndTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
- /*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
- */
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1654,7 +1525,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1683,6 +1554,148 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the
+ * current timeline to change. The record is already considered to
+ * be part of the new timeline, so we update ThisTimeLineID before
+ * replaying it. That's important so that replayEndTLI, which is
+ * recorded as the minimum recovery point's TLI if recovery stops
+ * after this record, is set correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newTLI = ThisTimeLineID;
+ TimeLineID prevTLI = ThisTimeLineID;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newTLI = checkPoint.ThisTimeLineID;
+ prevTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newTLI = xlrec.ThisTimeLineID;
+ prevTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newTLI != ThisTimeLineID)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
+
+ /* Following WAL records should be run with new TLI */
+ ThisTimeLineID = newTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so
+ * that XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecCtl->replayEndTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we
+ * see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the
+ * WAL record are consistent with the existing pages. This check
+ * is done only if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been
+ * successfully replayed.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up
+ * the receiver so that it notices the updated
+ * lastReplayedEndRecPtr and sends a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any
+ * (possibly bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new
+ * timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------915C630FC23A8D0E9CA95E65--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v5 3/3] Move code to apply one WAL record to a subroutine.
@ 2021-07-31 12:06 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-07-31 12:06 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 283 +++++++++++-----------
1 file changed, 148 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 6030d6fe819..85909c9b686 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -366,6 +366,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1374,11 +1375,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1400,11 +1398,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
(rmid == RM_XACT_ID && trace_recovery_messages <= DEBUG2) ||
@@ -1414,8 +1407,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1470,132 +1463,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
/*
- * ShmemVariableCache->nextXid must be beyond record's xid.
+ * Apply the record
*/
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newTLI = ThisTimeLineID;
- TimeLineID prevTLI = ThisTimeLineID;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newTLI = checkPoint.ThisTimeLineID;
- prevTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newTLI = xlrec.ThisTimeLineID;
- prevTLI = xlrec.PrevTimeLineID;
- }
-
- if (newTLI != ThisTimeLineID)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
-
- /* Following WAL records should be run with new TLI */
- ThisTimeLineID = newTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->replayEndRecPtr = EndRecPtr;
- XLogRecCtl->replayEndTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
- /*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
- */
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1654,7 +1525,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1683,6 +1554,148 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the
+ * current timeline to change. The record is already considered to
+ * be part of the new timeline, so we update ThisTimeLineID before
+ * replaying it. That's important so that replayEndTLI, which is
+ * recorded as the minimum recovery point's TLI if recovery stops
+ * after this record, is set correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newTLI = ThisTimeLineID;
+ TimeLineID prevTLI = ThisTimeLineID;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newTLI = checkPoint.ThisTimeLineID;
+ prevTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newTLI = xlrec.ThisTimeLineID;
+ prevTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newTLI != ThisTimeLineID)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
+
+ /* Following WAL records should be run with new TLI */
+ ThisTimeLineID = newTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so
+ * that XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecCtl->replayEndTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we
+ * see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the
+ * WAL record are consistent with the existing pages. This check
+ * is done only if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been
+ * successfully replayed.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up
+ * the receiver so that it notices the updated
+ * lastReplayedEndRecPtr and sends a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any
+ * (possibly bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new
+ * timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------F589ECF1004AF34308A0B206--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v6 3/3] Move code to apply one WAL record to a subroutine.
@ 2021-09-16 08:07 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-09-16 08:07 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 283 +++++++++++-----------
1 file changed, 148 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index e46215c2586..65edc7e1316 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -363,6 +363,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1361,11 +1362,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1387,11 +1385,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
(rmid == RM_XACT_ID && trace_recovery_messages <= DEBUG2) ||
@@ -1401,8 +1394,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1457,132 +1450,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
- /*
- * ShmemVariableCache->nextXid must be beyond record's xid.
- */
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newTLI = ThisTimeLineID;
- TimeLineID prevTLI = ThisTimeLineID;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newTLI = checkPoint.ThisTimeLineID;
- prevTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newTLI = xlrec.ThisTimeLineID;
- prevTLI = xlrec.PrevTimeLineID;
- }
-
- if (newTLI != ThisTimeLineID)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
-
- /* Following WAL records should be run with new TLI */
- ThisTimeLineID = newTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->replayEndRecPtr = EndRecPtr;
- XLogRecCtl->replayEndTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecCtl->info_lck);
- XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
- SpinLockRelease(&XLogRecCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
/*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
+ * Apply the record
*/
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1641,7 +1512,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1670,6 +1541,148 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the
+ * current timeline to change. The record is already considered to
+ * be part of the new timeline, so we update ThisTimeLineID before
+ * replaying it. That's important so that replayEndTLI, which is
+ * recorded as the minimum recovery point's TLI if recovery stops
+ * after this record, is set correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newTLI = ThisTimeLineID;
+ TimeLineID prevTLI = ThisTimeLineID;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newTLI = checkPoint.ThisTimeLineID;
+ prevTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newTLI = xlrec.ThisTimeLineID;
+ prevTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newTLI != ThisTimeLineID)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newTLI, prevTLI);
+
+ /* Following WAL records should be run with new TLI */
+ ThisTimeLineID = newTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so
+ * that XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecCtl->replayEndTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we
+ * see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the
+ * WAL record are consistent with the existing pages. This check
+ * is done only if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been
+ * successfully replayed.
+ */
+ SpinLockAcquire(&XLogRecCtl->info_lck);
+ XLogRecCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecCtl->lastReplayedTLI = ThisTimeLineID;
+ SpinLockRelease(&XLogRecCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up
+ * the receiver so that it notices the updated
+ * lastReplayedEndRecPtr and sends a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any
+ * (possibly bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, ThisTimeLineID);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new
+ * timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------7E727ABC31D380A53E56FBBC--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v8 4/4] Move code to apply one WAL record to a subroutine.
@ 2021-09-16 08:07 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-09-16 08:07 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 267 +++++++++++-----------
1 file changed, 139 insertions(+), 128 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index c21436190fd..72dc611567e 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -367,6 +367,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1398,7 +1399,6 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
@@ -1426,8 +1426,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
if (!StandbyMode)
ereport_startup_progress("redo in progress, elapsed time: %ld.%02d s, current LSN: %X/%X",
LSN_FORMAT_ARGS(xlogreader->ReadRecPtr));
@@ -1497,133 +1495,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
- /*
- * ShmemVariableCache->nextXid must be beyond record's xid.
- */
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newReplayTLI = replayTLI;
- TimeLineID prevReplayTLI = replayTLI;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newReplayTLI = checkPoint.ThisTimeLineID;
- prevReplayTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newReplayTLI = xlrec.ThisTimeLineID;
- prevReplayTLI = xlrec.PrevTimeLineID;
- }
-
- if (newReplayTLI != replayTLI)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(xlogreader->EndRecPtr, newReplayTLI,
- prevReplayTLI, replayTLI);
-
- /* Following WAL records should be run with new TLI */
- replayTLI = newReplayTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->replayEndRecPtr = xlogreader->EndRecPtr;
- XLogRecoveryCtl->replayEndTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->lastReplayedEndRecPtr = xlogreader->EndRecPtr;
- XLogRecoveryCtl->lastReplayedTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = xlogreader->ReadRecPtr;
-
/*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
+ * Apply the record
*/
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(xlogreader->EndRecPtr, replayTLI);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record, &replayTLI);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1711,6 +1586,142 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI)
+{
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the current
+ * timeline to change. The record is already considered to be part of the
+ * new timeline, so we update replayTLI before replaying it. That's
+ * important so that replayEndTLI, which is recorded as the minimum
+ * recovery point's TLI if recovery stops after this record, is set
+ * correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newReplayTLI = *replayTLI;
+ TimeLineID prevReplayTLI = *replayTLI;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newReplayTLI = checkPoint.ThisTimeLineID;
+ prevReplayTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newReplayTLI = xlrec.ThisTimeLineID;
+ prevReplayTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newReplayTLI != *replayTLI)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(xlogreader->EndRecPtr,
+ newReplayTLI, prevReplayTLI, *replayTLI);
+
+ /* Following WAL records should be run with new TLI */
+ *replayTLI = newReplayTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so that
+ * XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->replayEndRecPtr = xlogreader->EndRecPtr;
+ XLogRecoveryCtl->replayEndTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the WAL
+ * record are consistent with the existing pages. This check is done only
+ * if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been successfully
+ * replayed.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->lastReplayedEndRecPtr = xlogreader->EndRecPtr;
+ XLogRecoveryCtl->lastReplayedTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = xlogreader->ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up the
+ * receiver so that it notices the updated lastReplayedEndRecPtr and sends
+ * a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any (possibly
+ * bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(xlogreader->EndRecPtr, *replayTLI);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------2135AFA2DD79B96A7476E2B5--
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v7 5/5] Move code to apply one WAL record to a subroutine.
@ 2021-09-16 08:07 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-09-16 08:07 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 284 +++++++++++-----------
1 file changed, 147 insertions(+), 137 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 5b9d928a8ab..fe6b215b9c5 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -367,6 +367,7 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI);
static void xlog_block_info(StringInfo buf, XLogReaderState *record);
static void readRecoverySignalFile(void);
@@ -1396,11 +1397,8 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
- XLogRecPtr ReadRecPtr;
- XLogRecPtr EndRecPtr;
pg_rusage_init(&ru0);
@@ -1426,14 +1424,9 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
- ReadRecPtr = xlogreader->ReadRecPtr;
- EndRecPtr = xlogreader->EndRecPtr;
-
if (!StandbyMode)
ereport_startup_progress("redo in progress, elapsed time: %ld.%02d s, current LSN: %X/%X",
- LSN_FORMAT_ARGS(ReadRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr));
#ifdef WAL_DEBUG
if (XLOG_DEBUG ||
@@ -1444,8 +1437,8 @@ PerformWalRecovery(void)
initStringInfo(&buf);
appendStringInfo(&buf, "REDO @ %X/%X; LSN %X/%X: ",
- LSN_FORMAT_ARGS(ReadRecPtr),
- LSN_FORMAT_ARGS(EndRecPtr));
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->EndRecPtr));
xlog_outrec(&buf, xlogreader);
appendStringInfoString(&buf, " - ");
xlog_outdesc(&buf, xlogreader);
@@ -1500,133 +1493,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
- /*
- * ShmemVariableCache->nextXid must be beyond record's xid.
- */
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newReplayTLI = replayTLI;
- TimeLineID prevReplayTLI = replayTLI;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newReplayTLI = checkPoint.ThisTimeLineID;
- prevReplayTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newReplayTLI = xlrec.ThisTimeLineID;
- prevReplayTLI = xlrec.PrevTimeLineID;
- }
-
- if (newReplayTLI != replayTLI)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(EndRecPtr, newReplayTLI,
- prevReplayTLI, replayTLI);
-
- /* Following WAL records should be run with new TLI */
- replayTLI = newReplayTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->replayEndRecPtr = EndRecPtr;
- XLogRecoveryCtl->replayEndTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- checkXLogConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
- /*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
- */
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->lastReplayedEndRecPtr = EndRecPtr;
- XLogRecoveryCtl->lastReplayedTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = ReadRecPtr;
-
/*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
+ * Apply the record
*/
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(EndRecPtr, replayTLI);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record, &replayTLI);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1685,7 +1555,7 @@ PerformWalRecovery(void)
ereport(LOG,
(errmsg("redo done at %X/%X system usage: %s",
- LSN_FORMAT_ARGS(ReadRecPtr),
+ LSN_FORMAT_ARGS(xlogreader->ReadRecPtr),
pg_rusage_show(&ru0))));
xtime = GetLatestXTime();
if (xtime)
@@ -1714,6 +1584,146 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI)
+{
+ XLogRecPtr ReadRecPtr;
+ XLogRecPtr EndRecPtr;
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ ReadRecPtr = xlogreader->ReadRecPtr;
+ EndRecPtr = xlogreader->EndRecPtr;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the current
+ * timeline to change. The record is already considered to be part of the
+ * new timeline, so we update replayTLI before replaying it. That's
+ * important so that replayEndTLI, which is recorded as the minimum
+ * recovery point's TLI if recovery stops after this record, is set
+ * correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newReplayTLI = *replayTLI;
+ TimeLineID prevReplayTLI = *replayTLI;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newReplayTLI = checkPoint.ThisTimeLineID;
+ prevReplayTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newReplayTLI = xlrec.ThisTimeLineID;
+ prevReplayTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newReplayTLI != *replayTLI)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(EndRecPtr, newReplayTLI, prevReplayTLI, *replayTLI);
+
+ /* Following WAL records should be run with new TLI */
+ *replayTLI = newReplayTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so that
+ * XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->replayEndRecPtr = EndRecPtr;
+ XLogRecoveryCtl->replayEndTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the WAL
+ * record are consistent with the existing pages. This check is done only
+ * if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ checkXLogConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been successfully
+ * replayed.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->lastReplayedEndRecPtr = EndRecPtr;
+ XLogRecoveryCtl->lastReplayedTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up the
+ * receiver so that it notices the updated lastReplayedEndRecPtr and sends
+ * a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any (possibly
+ * bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(EndRecPtr, *replayTLI);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Error context callback for errors occurring during rm_redo().
*/
--
2.30.2
--------------3904592D294EBE853403D5C1--
^ permalink raw reply [nested|flat] 18+ messages in thread
* pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
@ 2021-10-28 01:50 Michael Paquier <[email protected]>
2022-04-11 07:43 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
0 siblings, 2 replies; 18+ messages in thread
From: Michael Paquier @ 2021-10-28 01:50 UTC (permalink / raw)
To: [email protected]
Add TAP test for archive_cleanup_command and recovery_end_command
This adds tests checking for the execution of both commands. The
recovery test 002_archiving.pl is nicely adapted to that, as promotion
is triggered already twice there, and even if any of those commands fail
they don't affect recovery or promotion.
A command success is checked using a file generated by an "echo"
command, that should be able to work in all the buildfarm environments,
even Msys (but we'll know soon about that). Command failure is tested
with an "echo" command that points to a path that does not exist,
scanning the backend logs to make sure that the failure happens. Both
rely on the backend triggering the commands from the root of the data
folder, making its logic more robust.
Thanks to Neha Sharma for the extra tests on Windows.
Author: Amul Sul, Michael Paquier
Reviewed-by: Andres Freund, Euler Taveira
Discussion: https://postgr.es/m/CAAJ_b95R_c4T5moq30qsybSU=eDzDHm=4SPiAWaiMWc2OW7=1Q@mail.gmail.com
Branch
------
master
Details
-------
https://git.postgresql.org/pg/commitdiff/46dea2419ee7895a4eb3d048317682e6f18a17e1
Modified Files
--------------
src/test/recovery/t/002_archiving.pl | 47 +++++++++++++++++++++++++++++++++++-
1 file changed, 46 insertions(+), 1 deletion(-)
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-11 07:43 ` Michael Paquier <[email protected]>
1 sibling, 0 replies; 18+ messages in thread
From: Michael Paquier @ 2022-04-11 07:43 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Apr 11, 2022 at 06:48:58PM +1200, Thomas Munro wrote:
> Sorry for the delay... I got a bit confused about the different things
> going on in this thread but I hope I've got it now:
>
> 1. This test had some pre-existing bugs/races, which hadn't failed
> before due to scheduling, even under Valgrind. The above changes
> appear to fix those problems. To Michael for comment.
I have seen the thread, and there is a lot in it. I will try to look
tomorrow at the parts I got involved in.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-12 03:49 ` Michael Paquier <[email protected]>
2022-04-17 04:17 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-18 04:55 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
1 sibling, 2 replies; 18+ messages in thread
From: Michael Paquier @ 2022-04-12 03:49 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Mon, Apr 11, 2022 at 06:48:58PM +1200, Thomas Munro wrote:
> 1. This test had some pre-existing bugs/races, which hadn't failed
> before due to scheduling, even under Valgrind. The above changes
> appear to fix those problems. To Michael for comment.
Yeah, there are two problems here. From what I can see, ensuring the
execution of archive_cleanup_command on the standby needs the
checkpoint on the primary and the restart point on the standby. So
pg_current_wal_lsn() should be located after the primary's checkpoint
and not before it so as we are sure that the checkpoint records finds
its way to the standby. That's what Tom mentioned upthread.
The second problem is to make sure that $standby2 sees the promotion
of $standby and its history file, but we also want to recover
00000002.history from some archives to create a RECOVERYHISTORY at
recovery for the purpose of the test. Switching to a new segment as
proposed by Andres does not seem completely right to me because we are
not 100% sure of the ordering an archive is going to happen, no? I
think that the logic to create $standby2 from the initial backup of
the primary is right, because there is no 00000002.history in it, but
we also need to be sure that 00000002.history has been archived once
the promotion of $standby is done. This can be validated thanks to
the logs, actually.
>> What is that second test really testing?
>>
>> # Check the presence of temporary files specifically generated during
>> # archive recovery. To ensure the presence of the temporary history
>> # file, switch to a timeline large enough to allow a standby to recover
>> # a history file from an archive. As this requires at least two timeline
>> # switches, promote the existing standby first. Then create a second
>> # standby based on the promoted one. Finally, the second standby is
>> # promoted.
>>
>> Note "Then create a second standby based on the promoted one." - but that's
>> not actually what's happening:
>
> 2. There may also be other problems with the test but those aren't
> relevant to skink's failure, which starts on the 5th test. To Michael
> for comment.
This comes from df86e52, where we want to recovery a history file that
would be created as RECOVERYHISTORY and make sure that the file gets
removed at the end of recovery. So $standby2 should choose a new
timeline different from the one of chosen by $standby. Looking back
at what has been done, it seems to me that the comment is the
incorrect part:
https://www.postgresql.org/message-id/[email protected]
All that stuff leads me to the attached. Thoughts?
--
Michael
Attachments:
[text/x-diff] tap-archiving-michael.patch (3.5K, ../../[email protected]/2-tap-archiving-michael.patch)
download | inline diff:
diff --git a/src/test/recovery/t/002_archiving.pl b/src/test/recovery/t/002_archiving.pl
index c8f5ffbaf0..45aafcb35c 100644
--- a/src/test/recovery/t/002_archiving.pl
+++ b/src/test/recovery/t/002_archiving.pl
@@ -24,6 +24,8 @@ $node_primary->backup($backup_name);
# Initialize standby node from backup, fetching WAL from archives
my $node_standby = PostgreSQL::Test::Cluster->new('standby');
+# Note that this makes the standby archive its contents on the archives
+# of the primary.
$node_standby->init_from_backup($node_primary, $backup_name,
has_restoring => 1);
$node_standby->append_conf('postgresql.conf',
@@ -44,13 +46,16 @@ $node_standby->start;
# Create some content on primary
$node_primary->safe_psql('postgres',
"CREATE TABLE tab_int AS SELECT generate_series(1,1000) AS a");
-my $current_lsn =
- $node_primary->safe_psql('postgres', "SELECT pg_current_wal_lsn();");
# Note the presence of this checkpoint for the archive_cleanup_command
# check done below, before switching to a new segment.
$node_primary->safe_psql('postgres', "CHECKPOINT");
+# Done after the checkpoint to ensure that the checkpoint gets replayed
+# on the standby.
+my $current_lsn =
+ $node_primary->safe_psql('postgres', "SELECT pg_current_wal_lsn();");
+
# Force archiving of WAL file to make it present on primary
$node_primary->safe_psql('postgres', "SELECT pg_switch_wal()");
@@ -81,10 +86,34 @@ ok( !-f "$data_dir/$recovery_end_command_file",
# file, switch to a timeline large enough to allow a standby to recover
# a history file from an archive. As this requires at least two timeline
# switches, promote the existing standby first. Then create a second
-# standby based on the promoted one. Finally, the second standby is
-# promoted.
+# standby based on the primary, using its archives. Finally, the second
+# standby is promoted.
$node_standby->promote;
+# Wait until the history file has been archived on the archives of the
+# primary once the promotion of the standby completes.
+my $primary_archive = $node_primary->archive_dir;
+my $max_attempts = 10 * $PostgreSQL::Test::Utils::timeout_default;
+my $attempts = 0;
+while ($attempts < $max_attempts)
+{
+ last if (-e "$primary_archive/00000002.history");
+
+ # Wait 0.1 second before retrying.
+ usleep(100_000);
+
+ $attempts++;
+}
+
+if ($attempts >= $max_attempts)
+{
+ die "timed out waiting for 00000002.history\n";
+}
+else
+{
+ note "found 00000002.history after $attempts attempts\n"
+}
+
# recovery_end_command should have been triggered on promotion.
ok( -f "$data_dir/$recovery_end_command_file",
'recovery_end_command executed after promotion');
@@ -108,14 +137,19 @@ my $log_location = -s $node_standby2->logfile;
# Now promote standby2, and check that temporary files specifically
# generated during archive recovery are removed by the end of recovery.
$node_standby2->promote;
+
+# Check the logs of the standby to see that the commands have failed.
+my $log_contents = slurp_file($node_standby2->logfile, $log_location);
my $node_standby2_data = $node_standby2->data_dir;
+
+like(
+ $log_contents,
+ qr/restored log file "00000002.history" from archive/s,
+ "00000002.history retrieved from the archives");
ok( !-f "$node_standby2_data/pg_wal/RECOVERYHISTORY",
"RECOVERYHISTORY removed after promotion");
ok( !-f "$node_standby2_data/pg_wal/RECOVERYXLOG",
"RECOVERYXLOG removed after promotion");
-
-# Check the logs of the standby to see that the commands have failed.
-my $log_contents = slurp_file($node_standby2->logfile, $log_location);
like(
$log_contents,
qr/WARNING:.*recovery_end_command/s,
[application/pgp-signature] signature.asc (833B, ../../[email protected]/3-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-17 04:17 ` Michael Paquier <[email protected]>
2022-04-17 23:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
1 sibling, 1 reply; 18+ messages in thread
From: Michael Paquier @ 2022-04-17 04:17 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Sun, Apr 17, 2022 at 08:56:33AM +1200, Thomas Munro wrote:
> Under valgrind I got "Undefined subroutine &main::usleep called at
> t/002_archiving.pl line 103" so I added "use Time::HiRes qw(usleep);",
> and now I get past the first 4 tests with your patch, but then
> promotion times out, not sure why:
>
> +++ tap check in src/test/recovery +++
> t/002_archiving.pl ..
> ok 1 - check content from archives
> ok 2 - archive_cleanup_command executed on checkpoint
> ok 3 - recovery_end_command not executed yet
> # found 00000002.history after 14 attempts
> ok 4 - recovery_end_command executed after promotion
> Bailout called. Further testing stopped: command "pg_ctl -D
> /home/tmunro/projects/postgresql/src/test/recovery/tmp_check/t_002_archiving_standby2_data/pgdata
> -l /home/tmunro/projects/postgresql/src/test/recovery/tmp_check/log/002_archiving_standby2.log
> promote" exited with value 1
Hmm. As far as I can see, aren't you just hitting the 60s timeout of
pg_ctl here due to the slowness of valgrind?
> Since it's quite painful to run TAP tests under valgrind, I found a
> place to stick a plain old sleep to repro these problems:
Actually, I am wondering how you are patching Cluster.pm to do that.
> Soon I'll push the fix to the slowness that xlogprefetcher.c
> accidentally introduced to continuous archive recovery, ie the problem
> of calling a failing restore_command repeatedly as we approach the end
> of a WAL segment instead of just once every 5 seconds after we run out
> of data, and after that you'll probably need to revert that fix
> locally to repro this.
Okay. Thanks. Anyway, I'll do something about that tomorrow (no
room to look at the buildfarm today), and I was thinking about
replacing the while loop I had in the last version of the patch with a
poll_query_until that does a pg_stat_file() with an absolute path to
the history file to avoid the dependency to usleep() in the test,
splitting the fix into two commits as there is more than one problem,
each applying to different branches.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-17 04:17 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-17 23:49 ` Michael Paquier <[email protected]>
2022-04-19 02:33 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
0 siblings, 1 reply; 18+ messages in thread
From: Michael Paquier @ 2022-04-17 23:49 UTC (permalink / raw)
To: Andrew Dunstan <[email protected]>; +Cc: Thomas Munro <[email protected]>; Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Sun, Apr 17, 2022 at 10:56:08AM -0400, Andrew Dunstan wrote:
> I don't really think it's Cluster.pm's business to deal with that. It
> takes an install path as given either explicitly or implicitly.
>
> It shouldn't be too hard to get Makefile.global to install valgrind
> wrappers into the tmp_install/bin directory.
Or what gets used in just a wrapper of the contents of bin/ that get
enforced to be first in PATH?
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../YlynfULh%[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-17 04:17 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-17 23:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-19 02:33 ` Michael Paquier <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Michael Paquier @ 2022-04-19 02:33 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Andrew Dunstan <[email protected]>; Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Apr 19, 2022 at 09:45:11AM +1200, Thomas Munro wrote:
> Delayed response to the question on how I did that, because it was a 4
> day weekend down here and I got distracted by sunshine...
Happy Easter.
> I think that sort of thing actually worked when I tried it on a
> beefier workstation, but it sent my Thinkpad that "only" has a 16GB of
> RAM into some kind of death spiral. The way I succeeded was indeed
> using a wrapper script, based on a suggestion from Andres, my
> kludgy-hardcoded-path-assuming implementation of which looked like:
>
> Yeah, it might be quite neat to find a tool-supported way to do that.
Thanks for the details. I feared that it was something like that for
the backend. At least that's better than having valgrind spawn all
the processes kicked by the make command. :/
> Tangentially, I'd also like to look into making
> PostgreSQL-under-Valgrind work on FreeBSD and macOS, which didn't work
> last time I tried it for reasons that might, I hope, have been fixed
> on the Valgrind side by now.
Okay.
As a side note, skink has cooled down since acf1dd4, and did not
complain either after the additions of e61efaf and 1a8b110.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
@ 2022-04-18 04:55 ` Michael Paquier <[email protected]>
1 sibling, 0 replies; 18+ messages in thread
From: Michael Paquier @ 2022-04-18 04:55 UTC (permalink / raw)
To: Thomas Munro <[email protected]>; +Cc: Andres Freund <[email protected]>; Tom Lane <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, Apr 12, 2022 at 12:49:48PM +0900, Michael Paquier wrote:
> This comes from df86e52, where we want to recovery a history file that
> would be created as RECOVERYHISTORY and make sure that the file gets
> removed at the end of recovery. So $standby2 should choose a new
> timeline different from the one of chosen by $standby. Looking back
> at what has been done, it seems to me that the comment is the
> incorrect part:
> https://www.postgresql.org/message-id/[email protected]
acf1dd42 has taken care of the failures of this test with skink, and I
have just taken care of the two races in the tests with e61efaf and
1a8b110. I have left e61efaf out of REL_10_STABLE as the idea of
relying on a poll_query_until() with pg_stat_file() and an absolute
path would not work there, and the branch will be EOL'd soon while
there were no complains with this test for two years.
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, ../../YlzvMeOnn%[email protected]/2-signature.asc)
download
^ permalink raw reply [nested|flat] 18+ messages in thread
* [PATCH v9 5/5] Move code to apply one WAL record to a subroutine.
@ 2021-12-17 07:00 Heikki Linnakangas <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Heikki Linnakangas @ 2021-12-17 07:00 UTC (permalink / raw)
---
src/backend/access/transam/xlogrecovery.c | 282 +++++++++++-----------
1 file changed, 147 insertions(+), 135 deletions(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index b9fb61d1dbb..875188cc7a8 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -374,6 +374,8 @@ static char recoveryStopName[MAXFNAMELEN];
static bool recoveryStopAfter;
/* prototypes for local functions */
+static void ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI);
+
static void readRecoverySignalFile(void);
static void validateRecoveryParameters(void);
static bool read_backup_label(XLogRecPtr *checkPointLoc,
@@ -1569,7 +1571,6 @@ PerformWalRecovery(void)
if (record != NULL)
{
- ErrorContextCallback errcallback;
TimestampTz xtime;
PGRUsage ru0;
@@ -1597,8 +1598,6 @@ PerformWalRecovery(void)
*/
do
{
- bool switchedTLI = false;
-
if (!StandbyMode)
ereport_startup_progress("redo in progress, elapsed time: %ld.%02d s, current LSN: %X/%X",
LSN_FORMAT_ARGS(xlogreader->ReadRecPtr));
@@ -1668,140 +1667,10 @@ PerformWalRecovery(void)
recoveryPausesHere(false);
}
- /* Setup error traceback support for ereport() */
- errcallback.callback = rm_redo_error_callback;
- errcallback.arg = (void *) xlogreader;
- errcallback.previous = error_context_stack;
- error_context_stack = &errcallback;
-
- /*
- * ShmemVariableCache->nextXid must be beyond record's xid.
- */
- AdvanceNextFullTransactionIdPastXid(record->xl_xid);
-
- /*
- * Before replaying this record, check if this record causes the
- * current timeline to change. The record is already considered to
- * be part of the new timeline, so we update ThisTimeLineID before
- * replaying it. That's important so that replayEndTLI, which is
- * recorded as the minimum recovery point's TLI if recovery stops
- * after this record, is set correctly.
- */
- if (record->xl_rmid == RM_XLOG_ID)
- {
- TimeLineID newReplayTLI = replayTLI;
- TimeLineID prevReplayTLI = replayTLI;
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
-
- if (info == XLOG_CHECKPOINT_SHUTDOWN)
- {
- CheckPoint checkPoint;
-
- memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
- newReplayTLI = checkPoint.ThisTimeLineID;
- prevReplayTLI = checkPoint.PrevTimeLineID;
- }
- else if (info == XLOG_END_OF_RECOVERY)
- {
- xl_end_of_recovery xlrec;
-
- memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
- newReplayTLI = xlrec.ThisTimeLineID;
- prevReplayTLI = xlrec.PrevTimeLineID;
- }
-
- if (newReplayTLI != replayTLI)
- {
- /* Check that it's OK to switch to this TLI */
- checkTimeLineSwitch(xlogreader->EndRecPtr, newReplayTLI,
- prevReplayTLI, replayTLI);
-
- /* Following WAL records should be run with new TLI */
- replayTLI = newReplayTLI;
- switchedTLI = true;
- }
- }
-
- /*
- * Update shared replayEndRecPtr before replaying this record, so
- * that XLogFlush will update minRecoveryPoint correctly.
- */
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->replayEndRecPtr = xlogreader->EndRecPtr;
- XLogRecoveryCtl->replayEndTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /*
- * If we are attempting to enter Hot Standby mode, process XIDs we
- * see
- */
- if (standbyState >= STANDBY_INITIALIZED &&
- TransactionIdIsValid(record->xl_xid))
- RecordKnownAssignedTransactionIds(record->xl_xid);
-
- /*
- * Some XLOG record types that are related to recovery are
- * processed directly here, rather than in xlog_redo()
- */
- if (record->xl_rmid == RM_XLOG_ID)
- xlogrecovery_redo(xlogreader, replayTLI);
-
- /* Now apply the WAL record itself */
- RmgrTable[record->xl_rmid].rm_redo(xlogreader);
-
- /*
- * After redo, check whether the backup pages associated with the
- * WAL record are consistent with the existing pages. This check
- * is done only if consistency check is enabled for this record.
- */
- if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
- verifyBackupPageConsistency(xlogreader);
-
- /* Pop the error context stack */
- error_context_stack = errcallback.previous;
-
/*
- * Update lastReplayedEndRecPtr after this record has been
- * successfully replayed.
+ * Apply the record
*/
- SpinLockAcquire(&XLogRecoveryCtl->info_lck);
- XLogRecoveryCtl->lastReplayedEndRecPtr = xlogreader->EndRecPtr;
- XLogRecoveryCtl->lastReplayedTLI = replayTLI;
- SpinLockRelease(&XLogRecoveryCtl->info_lck);
-
- /* Also remember its starting position. */
- LastReplayedReadRecPtr = xlogreader->ReadRecPtr;
-
- /*
- * If rm_redo called XLogRequestWalReceiverReply, then we wake up
- * the receiver so that it notices the updated
- * lastReplayedEndRecPtr and sends a reply to the primary.
- */
- if (doRequestWalReceiverReply)
- {
- doRequestWalReceiverReply = false;
- WalRcvForceReply();
- }
-
- /* Allow read-only connections if we're consistent now */
- CheckRecoveryConsistency();
-
- /* Is this a timeline switch? */
- if (switchedTLI)
- {
- /*
- * Before we continue on the new timeline, clean up any
- * (possibly bogus) future WAL segments on the old timeline.
- */
- RemoveNonParentXlogFiles(xlogreader->EndRecPtr, replayTLI);
-
- /*
- * Wake up any walsenders to notice that we are on a new
- * timeline.
- */
- if (AllowCascadeReplication())
- WalSndWakeup();
- }
+ ApplyWalRecord(xlogreader, record, &replayTLI);
/* Exit loop if we reached inclusive recovery target */
if (recoveryStopsAfter(xlogreader))
@@ -1889,6 +1758,149 @@ PerformWalRecovery(void)
(errmsg("recovery ended before configured recovery target was reached")));
}
+/*
+ * Subroutine of PerformWalRecovery, to apply one WAL record.
+ */
+static void
+ApplyWalRecord(XLogReaderState *xlogreader, XLogRecord *record, TimeLineID *replayTLI)
+{
+ ErrorContextCallback errcallback;
+ bool switchedTLI = false;
+
+ /* Setup error traceback support for ereport() */
+ errcallback.callback = rm_redo_error_callback;
+ errcallback.arg = (void *) xlogreader;
+ errcallback.previous = error_context_stack;
+ error_context_stack = &errcallback;
+
+ /*
+ * ShmemVariableCache->nextXid must be beyond record's xid.
+ */
+ AdvanceNextFullTransactionIdPastXid(record->xl_xid);
+
+ /*
+ * Before replaying this record, check if this record causes the current
+ * timeline to change. The record is already considered to be part of the
+ * new timeline, so we update replayTLI before replaying it. That's
+ * important so that replayEndTLI, which is recorded as the minimum
+ * recovery point's TLI if recovery stops after this record, is set
+ * correctly.
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ {
+ TimeLineID newReplayTLI = *replayTLI;
+ TimeLineID prevReplayTLI = *replayTLI;
+ uint8 info = record->xl_info & ~XLR_INFO_MASK;
+
+ if (info == XLOG_CHECKPOINT_SHUTDOWN)
+ {
+ CheckPoint checkPoint;
+
+ memcpy(&checkPoint, XLogRecGetData(xlogreader), sizeof(CheckPoint));
+ newReplayTLI = checkPoint.ThisTimeLineID;
+ prevReplayTLI = checkPoint.PrevTimeLineID;
+ }
+ else if (info == XLOG_END_OF_RECOVERY)
+ {
+ xl_end_of_recovery xlrec;
+
+ memcpy(&xlrec, XLogRecGetData(xlogreader), sizeof(xl_end_of_recovery));
+ newReplayTLI = xlrec.ThisTimeLineID;
+ prevReplayTLI = xlrec.PrevTimeLineID;
+ }
+
+ if (newReplayTLI != *replayTLI)
+ {
+ /* Check that it's OK to switch to this TLI */
+ checkTimeLineSwitch(xlogreader->EndRecPtr,
+ newReplayTLI, prevReplayTLI, *replayTLI);
+
+ /* Following WAL records should be run with new TLI */
+ *replayTLI = newReplayTLI;
+ switchedTLI = true;
+ }
+ }
+
+ /*
+ * Update shared replayEndRecPtr before replaying this record, so that
+ * XLogFlush will update minRecoveryPoint correctly.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->replayEndRecPtr = xlogreader->EndRecPtr;
+ XLogRecoveryCtl->replayEndTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /*
+ * If we are attempting to enter Hot Standby mode, process XIDs we see
+ */
+ if (standbyState >= STANDBY_INITIALIZED &&
+ TransactionIdIsValid(record->xl_xid))
+ RecordKnownAssignedTransactionIds(record->xl_xid);
+
+ /*
+ * Some XLOG record types that are related to recovery are processed
+ * directly here, rather than in xlog_redo()
+ */
+ if (record->xl_rmid == RM_XLOG_ID)
+ xlogrecovery_redo(xlogreader, *replayTLI);
+
+ /* Now apply the WAL record itself */
+ RmgrTable[record->xl_rmid].rm_redo(xlogreader);
+
+ /*
+ * After redo, check whether the backup pages associated with the WAL
+ * record are consistent with the existing pages. This check is done only
+ * if consistency check is enabled for this record.
+ */
+ if ((record->xl_info & XLR_CHECK_CONSISTENCY) != 0)
+ verifyBackupPageConsistency(xlogreader);
+
+ /* Pop the error context stack */
+ error_context_stack = errcallback.previous;
+
+ /*
+ * Update lastReplayedEndRecPtr after this record has been successfully
+ * replayed.
+ */
+ SpinLockAcquire(&XLogRecoveryCtl->info_lck);
+ XLogRecoveryCtl->lastReplayedEndRecPtr = xlogreader->EndRecPtr;
+ XLogRecoveryCtl->lastReplayedTLI = *replayTLI;
+ SpinLockRelease(&XLogRecoveryCtl->info_lck);
+
+ /* Also remember its starting position. */
+ LastReplayedReadRecPtr = xlogreader->ReadRecPtr;
+
+ /*
+ * If rm_redo called XLogRequestWalReceiverReply, then we wake up the
+ * receiver so that it notices the updated lastReplayedEndRecPtr and sends
+ * a reply to the primary.
+ */
+ if (doRequestWalReceiverReply)
+ {
+ doRequestWalReceiverReply = false;
+ WalRcvForceReply();
+ }
+
+ /* Allow read-only connections if we're consistent now */
+ CheckRecoveryConsistency();
+
+ /* Is this a timeline switch? */
+ if (switchedTLI)
+ {
+ /*
+ * Before we continue on the new timeline, clean up any (possibly
+ * bogus) future WAL segments on the old timeline.
+ */
+ RemoveNonParentXlogFiles(xlogreader->EndRecPtr, *replayTLI);
+
+ /*
+ * Wake up any walsenders to notice that we are on a new timeline.
+ */
+ if (AllowCascadeReplication())
+ WalSndWakeup();
+ }
+}
+
/*
* Some XLOG RM record types that are directly related to WAL recovery are
* handled here rather than in the xlog_redo()
--
2.30.2
--------------24AFD7B565D1554637A9E916--
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row
@ 2024-11-28 15:33 Kirill Reshke <[email protected]>
2024-12-13 03:15 ` Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row jian he <[email protected]>
0 siblings, 1 reply; 18+ messages in thread
From: Kirill Reshke @ 2024-11-28 15:33 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Fujii Masao <[email protected]>; Jim Jones <[email protected]>; David G. Johnston <[email protected]>; Yugo NAGATA <[email protected]>; torikoshia <[email protected]>; PostgreSQL Hackers <[email protected]>
On Tue, 19 Nov 2024 at 13:52, jian he <[email protected]> wrote:
>
> On Sat, Nov 16, 2024 at 5:55 PM Kirill Reshke <[email protected]> wrote:
> >
> > I am attaching my v8 for reference.
> >
>
> in your v8.
>
> <varlistentry>
> <term><literal>REJECT_LIMIT</literal></term>
> <listitem>
> <para>
> Specifies the maximum number of errors tolerated while converting a
> column's input value to its data type, when <literal>ON_ERROR</literal> is
> set to <literal>ignore</literal>.
> If the input contains more erroneous rows than the specified
> value, the <command>COPY</command>
> command fails, even with <literal>ON_ERROR</literal> set to
> <literal>ignore</literal>.
> </para>
> </listitem>
> </varlistentry>
>
> then above description not meet with following example, (i think)
>
> create table t(a int not null);
> COPY t FROM STDIN WITH (on_error set_to_null, reject_limit 2);
> Enter data to be copied followed by a newline.
> End with a backslash and a period on a line by itself, or an EOF signal.
> >> a
> >> \.
> ERROR: null value in column "a" of relation "t" violates not-null constraint
> DETAIL: Failing row contains (null).
> CONTEXT: COPY t, line 1, column a: "a"
>
> Overall, I think
> making the domain not-null align with column level not-null would be a
> good thing.
>
>
> <para>
> Specifies how to behave when encountering an error converting a column's
> input value into its data type.
> An <replaceable class="parameter">error_action</replaceable> value of
> <literal>stop</literal> means fail the command,
> <literal>ignore</literal> means discard the input row and
> continue with the next one, and
> <literal>set_to_null</literal> means replace columns containing
> erroneous input values with <literal>null</literal> and move to the
> next row.
>
> "and move to the next row" is wrong?
> I think it should be " and move to the next field".
Hi! There is not too much time left in this CF, so I moved to the next one.
If you are going to work on this patch, I'm waiting on your feedback
or a v9 patch that answers the issues brought up in this discussion.
--
Best regards,
Kirill Reshke
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row
2024-11-28 15:33 Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row Kirill Reshke <[email protected]>
@ 2024-12-13 03:15 ` jian he <[email protected]>
2025-01-08 07:05 ` Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row Kirill Reshke <[email protected]>
0 siblings, 1 reply; 18+ messages in thread
From: jian he @ 2024-12-13 03:15 UTC (permalink / raw)
To: Kirill Reshke <[email protected]>; +Cc: Fujii Masao <[email protected]>; Jim Jones <[email protected]>; David G. Johnston <[email protected]>; Yugo NAGATA <[email protected]>; torikoshia <[email protected]>; PostgreSQL Hackers <[email protected]>
+ /*
+ * Here we end processing of current COPY row.
+ * Update copy state counter for number of erroneous rows.
+ */
+ cstate->num_errors++;
+ cstate->escontext->error_occurred = true;
+
+ /* Only print this NOTICE message, if it will not be followed by ERROR */
+ if (cstate->opts.log_verbosity == COPY_LOG_VERBOSITY_VERBOSE &&
+ (
+ (cstate->opts.on_error == COPY_ON_ERROR_NULL &&
cstate->opts.reject_limit > 0 && cstate->num_errors <=
cstate->opts.reject_limit) ||
+ (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
(cstate->opts.reject_limit == 0 || cstate->num_errors <=
cstate->opts.reject_limit))
+ ))
{
this is kind of hard to comprehend.
so attached is a simple version of it based on v8.
for copy (on_error set_to_null)
1. not allow specifying reject_limit option
2. ereport ERROR for not-null constraint violation for domain type.
for example:
CREATE DOMAIN d_int_not_null AS INT NOT NULL CHECK(value > 0);
CREATE TABLE t1 (a d_int_not_null);
COPY t1 FROM STDIN WITH (on_error set_to_null);
these 3 values: \N a -1
will error out, the error message will be:
ERROR: domain d_int_not_null does not allow null values
Attachments:
[text/x-patch] v9-0001-new-COPY-on_error-option-set_to_null.patch (19.4K, ../../CACJufxEKq25CvQj4rYxcvHC_0MZQxzfwEKVWo9i-bKKxbVw3BA@mail.gmail.com/2-v9-0001-new-COPY-on_error-option-set_to_null.patch)
download | inline diff:
From 47bdeda8911596950463e70e33253e773ef13192 Mon Sep 17 00:00:00 2001
From: jian he <[email protected]>
Date: Fri, 13 Dec 2024 11:03:16 +0800
Subject: [PATCH v9 1/1] new COPY on_error option: set_to_null
extent "on_error action", introduce new option: on_error set_to_null.
due to current grammar, we cannot use "on_error null", so i choose on_error set_to_null.
any data type conversion errors while the COPY FROM process will set that column value to be NULL.
this will only work with COPY FROM with non-binary format.
However this will respect the not-null constraint, meaning, if you actually
converted error to null, but the column has not-null constraint, not-null
constraint violation ERROR will be reported.
regress test contains corner case for not-null with domain constraint.
discussion: https://postgr.es/m/CAKFQuwawy1e6YR4S=j+y7pXqg_Dw1WBVrgvf=BP3d1_aSfe_+Q@mail.gmail.com
---
doc/src/sgml/ref/copy.sgml | 26 ++++++++----
src/backend/commands/copy.c | 6 ++-
src/backend/commands/copyfrom.c | 31 +++++++++------
src/backend/commands/copyfromparse.c | 47 +++++++++++++++++++++-
src/bin/psql/tab-complete.in.c | 2 +-
src/include/commands/copy.h | 1 +
src/include/commands/copyfrom_internal.h | 4 +-
src/test/regress/expected/copy2.out | 50 ++++++++++++++++++++++++
src/test/regress/sql/copy2.sql | 45 +++++++++++++++++++++
9 files changed, 188 insertions(+), 24 deletions(-)
diff --git a/doc/src/sgml/ref/copy.sgml b/doc/src/sgml/ref/copy.sgml
index 8394402f09..4b847c25f3 100644
--- a/doc/src/sgml/ref/copy.sgml
+++ b/doc/src/sgml/ref/copy.sgml
@@ -394,21 +394,33 @@ COPY { <replaceable class="parameter">table_name</replaceable> [ ( <replaceable
Specifies how to behave when encountering an error converting a column's
input value into its data type.
An <replaceable class="parameter">error_action</replaceable> value of
- <literal>stop</literal> means fail the command, while
- <literal>ignore</literal> means discard the input row and continue with the next one.
+ <literal>stop</literal> means fail the command,
+ <literal>ignore</literal> means discard the input row and continue with the next one, and
+ <literal>set_to_null</literal> means replace columns containing erroneous input values with <literal>null</literal> and move to the next field.
The default is <literal>stop</literal>.
</para>
<para>
- The <literal>ignore</literal> option is applicable only for <command>COPY FROM</command>
+ The <literal>ignore</literal> and <literal>set_to_null</literal> options are applicable only for <command>COPY FROM</command>
when the <literal>FORMAT</literal> is <literal>text</literal> or <literal>csv</literal>.
</para>
<para>
- A <literal>NOTICE</literal> message containing the ignored row count is
+ For <literal>ignore</literal> option,
+ a <literal>NOTICE</literal> message containing the ignored row count is
emitted at the end of the <command>COPY FROM</command> if at least one
- row was discarded. When <literal>LOG_VERBOSITY</literal> option is set to
- <literal>verbose</literal>, a <literal>NOTICE</literal> message
+ row was discarded.
+ For <literal>set_to_null</literal> option,
+ a <literal>NOTICE</literal> message containing the row count that erroneous input values replaced by to null happened is
+ emitted at the end of the <command>COPY FROM</command> if at least one row was replaced.
+ </para>
+ <para>
+ When <literal>LOG_VERBOSITY</literal> option is set to
+ <literal>verbose</literal>, for <literal>ignore</literal> option, a <literal>NOTICE</literal> message
containing the line of the input file and the column name whose input
- conversion has failed is emitted for each discarded row.
+ conversion has failed is emitted for each discarded row;
+ for <literal>set_to_null</literal> option,
+ a <literal>NOTICE</literal> message
+ containing the line of the input file and the column name whose input value has been replaced by null
+ is emitted for each row where input conversion has failed ;
When it is set to <literal>silent</literal>, no message is emitted
regarding ignored rows.
</para>
diff --git a/src/backend/commands/copy.c b/src/backend/commands/copy.c
index 2d98ecf3f4..f494d2d64c 100644
--- a/src/backend/commands/copy.c
+++ b/src/backend/commands/copy.c
@@ -403,12 +403,14 @@ defGetCopyOnErrorChoice(DefElem *def, ParseState *pstate, bool is_from)
parser_errposition(pstate, def->location)));
/*
- * Allow "stop", or "ignore" values.
+ * Allow "stop", "ignore", "set_to_null" values.
*/
if (pg_strcasecmp(sval, "stop") == 0)
return COPY_ON_ERROR_STOP;
if (pg_strcasecmp(sval, "ignore") == 0)
return COPY_ON_ERROR_IGNORE;
+ if (pg_strcasecmp(sval, "set_to_null") == 0)
+ return COPY_ON_ERROR_NULL;
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
@@ -918,7 +920,7 @@ ProcessCopyOptions(ParseState *pstate,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("only ON_ERROR STOP is allowed in BINARY mode")));
- if (opts_out->reject_limit && !opts_out->on_error)
+ if (opts_out->reject_limit && opts_out->on_error != COPY_ON_ERROR_IGNORE)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
/*- translator: first and second %s are the names of COPY option, e.g.
diff --git a/src/backend/commands/copyfrom.c b/src/backend/commands/copyfrom.c
index 4d52c93c30..a9e61f5e26 100644
--- a/src/backend/commands/copyfrom.c
+++ b/src/backend/commands/copyfrom.c
@@ -1003,7 +1003,7 @@ CopyFrom(CopyFromState cstate)
if (!NextCopyFrom(cstate, econtext, myslot->tts_values, myslot->tts_isnull))
break;
- if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
+ if ((cstate->opts.on_error == COPY_ON_ERROR_IGNORE || cstate->opts.on_error == COPY_ON_ERROR_NULL) &&
cstate->escontext->error_occurred)
{
/*
@@ -1321,14 +1321,22 @@ CopyFrom(CopyFromState cstate)
/* Done, clean up */
error_context_stack = errcallback.previous;
- if (cstate->opts.on_error != COPY_ON_ERROR_STOP &&
- cstate->num_errors > 0 &&
+ if (cstate->num_errors > 0 &&
cstate->opts.log_verbosity >= COPY_LOG_VERBOSITY_DEFAULT)
- ereport(NOTICE,
- errmsg_plural("%llu row was skipped due to data type incompatibility",
- "%llu rows were skipped due to data type incompatibility",
- (unsigned long long) cstate->num_errors,
- (unsigned long long) cstate->num_errors));
+ {
+ if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE)
+ ereport(NOTICE,
+ errmsg_plural("%llu row was skipped due to data type incompatibility",
+ "%llu rows were skipped due to data type incompatibility",
+ (unsigned long long) cstate->num_errors,
+ (unsigned long long) cstate->num_errors));
+ else if (cstate->opts.on_error == COPY_ON_ERROR_NULL)
+ ereport(NOTICE,
+ errmsg_plural("Erroneous values in %llu row was replaced with NULL",
+ "Erroneous values in %llu rows were replaced with NULL",
+ (unsigned long long) cstate->num_errors,
+ (unsigned long long) cstate->num_errors));
+ }
if (bistate != NULL)
FreeBulkInsertState(bistate);
@@ -1474,10 +1482,11 @@ BeginCopyFrom(ParseState *pstate,
cstate->escontext->error_occurred = false;
/*
- * Currently we only support COPY_ON_ERROR_IGNORE. We'll add other
- * options later
+ * Currently we only support COPY_ON_ERROR_IGNORE, COPY_ON_ERROR_NULL.
+ * We'll add other options later
*/
- if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE)
+ if (cstate->opts.on_error == COPY_ON_ERROR_IGNORE ||
+ cstate->opts.on_error == COPY_ON_ERROR_NULL)
cstate->escontext->details_wanted = false;
}
else
diff --git a/src/backend/commands/copyfromparse.c b/src/backend/commands/copyfromparse.c
index d1d43b53d8..7f98f66ebe 100644
--- a/src/backend/commands/copyfromparse.c
+++ b/src/backend/commands/copyfromparse.c
@@ -871,6 +871,7 @@ NextCopyFrom(CopyFromState cstate, ExprContext *econtext,
int fldct;
int fieldno;
char *string;
+ bool current_row_erroneous = false;
/* read raw fields in the next line */
if (!NextCopyFromRawFields(cstate, &field_strings, &fldct))
@@ -949,7 +950,8 @@ NextCopyFrom(CopyFromState cstate, ExprContext *econtext,
/*
* If ON_ERROR is specified with IGNORE, skip rows with soft
- * errors
+ * errors. If ON_ERROR is specified with SET_TO_NULL, try
+ * to replace attribute value with NULL.
*/
else if (!InputFunctionCallSafe(&in_functions[m],
string,
@@ -960,9 +962,48 @@ NextCopyFrom(CopyFromState cstate, ExprContext *econtext,
{
Assert(cstate->opts.on_error != COPY_ON_ERROR_STOP);
+ /*
+ * we use it to count the number of rows (not fields)
+ * successfully doing on_error set_to_null.
+ */
+ current_row_erroneous = true;
+
+ if (cstate->opts.on_error == COPY_ON_ERROR_NULL)
+ {
+ /*
+ * we need another InputFunctionCallSafe for domain with a
+ * not-null constraint. if domain don't have not-null
+ * then continue to the next field, otherwise error out.
+ */
+ cstate->escontext->error_occurred = false;
+ if (InputFunctionCallSafe(&in_functions[m],
+ NULL,
+ typioparams[m],
+ att->atttypmod,
+ (Node *) cstate->escontext,
+ &values[m]))
+ {
+ nulls[m] = true;
+
+ if (cstate->opts.log_verbosity == COPY_LOG_VERBOSITY_VERBOSE)
+ ereport(NOTICE,
+ errmsg("replaced row attribute \"%s\" with NULL due to data type incompatibility at line %llu",
+ cstate->cur_attname,
+ (unsigned long long) cstate->cur_lineno));
+ continue;
+ }
+ else
+ ereport(ERROR,
+ errcode(ERRCODE_NOT_NULL_VIOLATION),
+ errmsg("domain %s does not allow null values", format_type_be(typioparams[m])),
+ errdatatype(typioparams[m]));
+ }
+
cstate->num_errors++;
+ cstate->escontext->error_occurred = true;
- if (cstate->opts.log_verbosity == COPY_LOG_VERBOSITY_VERBOSE)
+ if (cstate->opts.log_verbosity == COPY_LOG_VERBOSITY_VERBOSE &&
+ cstate->opts.on_error != COPY_ON_ERROR_NULL)
{
/*
* Since we emit line number and column info in the below
@@ -1001,6 +1042,8 @@ NextCopyFrom(CopyFromState cstate, ExprContext *econtext,
cstate->cur_attval = NULL;
}
+ if (current_row_erroneous)
+ cstate->num_errors++;
Assert(fieldno == attr_count);
}
else
diff --git a/src/bin/psql/tab-complete.in.c b/src/bin/psql/tab-complete.in.c
index bbd08770c3..fc2260b541 100644
--- a/src/bin/psql/tab-complete.in.c
+++ b/src/bin/psql/tab-complete.in.c
@@ -3235,7 +3235,7 @@ match_previous_words(int pattern_id,
COMPLETE_WITH("FORMAT", "FREEZE", "DELIMITER", "NULL",
"HEADER", "QUOTE", "ESCAPE", "FORCE_QUOTE",
"FORCE_NOT_NULL", "FORCE_NULL", "ENCODING", "DEFAULT",
- "ON_ERROR", "LOG_VERBOSITY");
+ "ON_ERROR", "SET_TO_NULL", "LOG_VERBOSITY");
/* Complete COPY <sth> FROM|TO filename WITH (FORMAT */
else if (Matches("COPY|\\copy", MatchAny, "FROM|TO", MatchAny, "WITH", "(", "FORMAT"))
diff --git a/src/include/commands/copy.h b/src/include/commands/copy.h
index 4002a7f538..051ca12d10 100644
--- a/src/include/commands/copy.h
+++ b/src/include/commands/copy.h
@@ -38,6 +38,7 @@ typedef enum CopyOnErrorChoice
{
COPY_ON_ERROR_STOP = 0, /* immediately throw errors, default */
COPY_ON_ERROR_IGNORE, /* ignore errors */
+ COPY_ON_ERROR_NULL, /* set error field to null */
} CopyOnErrorChoice;
/*
diff --git a/src/include/commands/copyfrom_internal.h b/src/include/commands/copyfrom_internal.h
index cad52fcc78..5eaacd8c6b 100644
--- a/src/include/commands/copyfrom_internal.h
+++ b/src/include/commands/copyfrom_internal.h
@@ -98,7 +98,9 @@ typedef struct CopyFromStateData
ErrorSaveContext *escontext; /* soft error trapper during in_functions
* execution */
uint64 num_errors; /* total number of rows which contained soft
- * errors */
+ * errors, for ON_ERROR set_to_null, it's the
+ * number of rows successfully converted to null
+ */
int *defmap; /* array of default att numbers related to
* missing att */
ExprState **defexprs; /* array of default att expressions for all
diff --git a/src/test/regress/expected/copy2.out b/src/test/regress/expected/copy2.out
index 64ea33aeae..f03dd24d54 100644
--- a/src/test/regress/expected/copy2.out
+++ b/src/test/regress/expected/copy2.out
@@ -81,6 +81,12 @@ COPY x from stdin (on_error ignore, on_error ignore);
ERROR: conflicting or redundant options
LINE 1: COPY x from stdin (on_error ignore, on_error ignore);
^
+COPY x from stdin (on_error set_to_null, on_error set_to_null);
+ERROR: conflicting or redundant options
+LINE 1: COPY x from stdin (on_error set_to_null, on_error set_to_nul...
+ ^
+COPY x FROM stdin WITH (on_error set_to_null, reject_limit 2);
+ERROR: COPY REJECT_LIMIT requires ON_ERROR to be set to IGNORE
COPY x from stdin (log_verbosity default, log_verbosity verbose);
ERROR: conflicting or redundant options
LINE 1: COPY x from stdin (log_verbosity default, log_verbosity verb...
@@ -92,6 +98,8 @@ COPY x from stdin (format BINARY, null 'x');
ERROR: cannot specify NULL in BINARY mode
COPY x from stdin (format BINARY, on_error ignore);
ERROR: only ON_ERROR STOP is allowed in BINARY mode
+COPY x from stdin (format BINARY, on_error set_to_null);
+ERROR: only ON_ERROR STOP is allowed in BINARY mode
COPY x from stdin (on_error unsupported);
ERROR: COPY ON_ERROR "unsupported" not recognized
LINE 1: COPY x from stdin (on_error unsupported);
@@ -124,6 +132,10 @@ COPY x to stdout (format BINARY, on_error unsupported);
ERROR: COPY ON_ERROR cannot be used with COPY TO
LINE 1: COPY x to stdout (format BINARY, on_error unsupported);
^
+COPY x to stdin (on_error set_to_null);
+ERROR: COPY ON_ERROR cannot be used with COPY TO
+LINE 1: COPY x to stdin (on_error set_to_null);
+ ^
COPY x from stdin (log_verbosity unsupported);
ERROR: COPY LOG_VERBOSITY "unsupported" not recognized
LINE 1: COPY x from stdin (log_verbosity unsupported);
@@ -769,6 +781,42 @@ CONTEXT: COPY check_ign_err
NOTICE: skipping row due to data type incompatibility at line 8 for column "k": "a"
CONTEXT: COPY check_ign_err
NOTICE: 6 rows were skipped due to data type incompatibility
+CREATE DOMAIN d_int_not_null AS INT NOT NULL CHECK(value > 0);
+CREATE DOMAIN d_int_positive_maybe_null AS INT CHECK(value > 0);
+CREATE TABLE t_on_error_null (a d_int_not_null, b d_int_positive_maybe_null, c INT);
+\pset null NULL
+--fail, colum a cannot set to NULL value
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+ERROR: domain d_int_not_null does not allow null values
+CONTEXT: COPY t_on_error_null, line 1, column a: null input
+--fail, colum a is domain with not-null constraint
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+ERROR: domain d_int_not_null does not allow null values
+CONTEXT: COPY t_on_error_null, line 1, column a: "a"
+--fail, colum a cannot set to NULL value
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+ERROR: domain d_int_not_null does not allow null values
+CONTEXT: COPY t_on_error_null, line 1, column a: "-1"
+--fail. less data
+COPY t_on_error_null FROM STDIN WITH (delimiter ',', on_error set_to_null);
+ERROR: missing data for column "c"
+CONTEXT: COPY t_on_error_null, line 1: "1,1"
+--fail. extra data
+COPY t_on_error_null FROM STDIN WITH (delimiter ',', on_error set_to_null);
+ERROR: extra data after last expected column
+CONTEXT: COPY t_on_error_null, line 1: "1,2,3,4"
+--ok
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+NOTICE: Erroneous values in 3 rows were replaced with NULL
+-- check inserted content
+select * from t_on_error_null;
+ a | b | c
+----+------+------
+ 10 | NULL | NULL
+ 11 | NULL | 12
+ 13 | 14 | NULL
+(3 rows)
+
-- tests for on_error option with log_verbosity and null constraint via domain
CREATE DOMAIN dcheck_ign_err2 varchar(15) NOT NULL;
CREATE TABLE check_ign_err2 (n int, m int[], k int, l dcheck_ign_err2);
@@ -828,6 +876,8 @@ DROP VIEW instead_of_insert_tbl_view;
DROP VIEW instead_of_insert_tbl_view_2;
DROP FUNCTION fun_instead_of_insert_tbl();
DROP TABLE check_ign_err;
+DROP TABLE t_on_error_null;
+DROP DOMAIN d_int_not_null;
DROP TABLE check_ign_err2;
DROP DOMAIN dcheck_ign_err2;
DROP TABLE hard_err;
diff --git a/src/test/regress/sql/copy2.sql b/src/test/regress/sql/copy2.sql
index 45273557ce..2bbb1ecce1 100644
--- a/src/test/regress/sql/copy2.sql
+++ b/src/test/regress/sql/copy2.sql
@@ -67,12 +67,15 @@ COPY x from stdin (force_null (a), force_null (b));
COPY x from stdin (convert_selectively (a), convert_selectively (b));
COPY x from stdin (encoding 'sql_ascii', encoding 'sql_ascii');
COPY x from stdin (on_error ignore, on_error ignore);
+COPY x from stdin (on_error set_to_null, on_error set_to_null);
+COPY x FROM stdin WITH (on_error set_to_null, reject_limit 2);
COPY x from stdin (log_verbosity default, log_verbosity verbose);
-- incorrect options
COPY x from stdin (format BINARY, delimiter ',');
COPY x from stdin (format BINARY, null 'x');
COPY x from stdin (format BINARY, on_error ignore);
+COPY x from stdin (format BINARY, on_error set_to_null);
COPY x from stdin (on_error unsupported);
COPY x from stdin (format TEXT, force_quote(a));
COPY x from stdin (format TEXT, force_quote *);
@@ -87,6 +90,7 @@ COPY x from stdin (format TEXT, force_null *);
COPY x to stdout (format CSV, force_null(a));
COPY x to stdout (format CSV, force_null *);
COPY x to stdout (format BINARY, on_error unsupported);
+COPY x to stdin (on_error set_to_null);
COPY x from stdin (log_verbosity unsupported);
COPY x from stdin with (reject_limit 1);
COPY x from stdin with (on_error ignore, reject_limit 0);
@@ -534,6 +538,45 @@ a {2} 2
8 {8} 8
\.
+CREATE DOMAIN d_int_not_null AS INT NOT NULL CHECK(value > 0);
+CREATE DOMAIN d_int_positive_maybe_null AS INT CHECK(value > 0);
+CREATE TABLE t_on_error_null (a d_int_not_null, b d_int_positive_maybe_null, c INT);
+
+\pset null NULL
+--fail, colum a cannot set to NULL value
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+\N 11 13
+\.
+
+--fail, colum a is domain with not-null constraint
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+a 11 14
+\.
+
+--fail, colum a cannot set to NULL value
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+-1 11 13
+\.
+
+--fail. less data
+COPY t_on_error_null FROM STDIN WITH (delimiter ',', on_error set_to_null);
+1,1
+\.
+--fail. extra data
+COPY t_on_error_null FROM STDIN WITH (delimiter ',', on_error set_to_null);
+1,2,3,4
+\.
+
+--ok
+COPY t_on_error_null FROM STDIN WITH (on_error set_to_null);
+10 a d
+11 b 12
+13 14 e
+\.
+
+-- check inserted content
+select * from t_on_error_null;
+
-- tests for on_error option with log_verbosity and null constraint via domain
CREATE DOMAIN dcheck_ign_err2 varchar(15) NOT NULL;
CREATE TABLE check_ign_err2 (n int, m int[], k int, l dcheck_ign_err2);
@@ -603,6 +646,8 @@ DROP VIEW instead_of_insert_tbl_view;
DROP VIEW instead_of_insert_tbl_view_2;
DROP FUNCTION fun_instead_of_insert_tbl();
DROP TABLE check_ign_err;
+DROP TABLE t_on_error_null;
+DROP DOMAIN d_int_not_null;
DROP TABLE check_ign_err2;
DROP DOMAIN dcheck_ign_err2;
DROP TABLE hard_err;
--
2.34.1
^ permalink raw reply [nested|flat] 18+ messages in thread
* Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row
2024-11-28 15:33 Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row Kirill Reshke <[email protected]>
2024-12-13 03:15 ` Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row jian he <[email protected]>
@ 2025-01-08 07:05 ` Kirill Reshke <[email protected]>
0 siblings, 0 replies; 18+ messages in thread
From: Kirill Reshke @ 2025-01-08 07:05 UTC (permalink / raw)
To: jian he <[email protected]>; +Cc: Fujii Masao <[email protected]>; Jim Jones <[email protected]>; David G. Johnston <[email protected]>; Yugo NAGATA <[email protected]>; torikoshia <[email protected]>; PostgreSQL Hackers <[email protected]>
Thank you for your update.
On Fri, 13 Dec 2024 at 08:15, jian he <[email protected]> wrote:
>
> + /*
> + * Here we end processing of current COPY row.
> + * Update copy state counter for number of erroneous rows.
> + */
> + cstate->num_errors++;
> + cstate->escontext->error_occurred = true;
> +
> + /* Only print this NOTICE message, if it will not be followed by ERROR */
> + if (cstate->opts.log_verbosity == COPY_LOG_VERBOSITY_VERBOSE &&
> + (
> + (cstate->opts.on_error == COPY_ON_ERROR_NULL &&
> cstate->opts.reject_limit > 0 && cstate->num_errors <=
> cstate->opts.reject_limit) ||
> + (cstate->opts.on_error == COPY_ON_ERROR_IGNORE &&
> (cstate->opts.reject_limit == 0 || cstate->num_errors <=
> cstate->opts.reject_limit))
> + ))
> {
> this is kind of hard to comprehend.
> so attached is a simple version of it based on v8.
+1
So, in this version you essentially removed support for REJECT_LIMIT +
SET_TO_NULL feature? Looks like a promising change. It is more likely
to see this committed.
So, +1 on that too.
However, v9 lacks tests for REJECT_LIMIT vs erroneous rows tests.
In short, we need this message somewhere in a regression test.
```
ERROR: skipped more than REJECT_LIMIT (xxx) rows due to data type
incompatibility
```
Also, please update commit msg with all authors and reviewers. This
will make committer job a little bit easier
--
Best regards,
Kirill Reshke
^ permalink raw reply [nested|flat] 18+ messages in thread
end of thread, other threads:[~2025-01-08 07:05 UTC | newest]
Thread overview: 18+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2021-06-21 21:00 [PATCH 7/7] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-06-21 21:00 [PATCH 7/7] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-07-31 12:06 [PATCH v4 3/3] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-07-31 12:06 [PATCH v5 3/3] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-09-16 08:07 [PATCH v6 3/3] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-09-16 08:07 [PATCH v8 4/4] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-09-16 08:07 [PATCH v7 5/5] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2021-10-28 01:50 pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-11 07:43 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-12 03:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-17 04:17 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-17 23:49 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-19 02:33 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2022-04-18 04:55 ` Re: pgsql: Add TAP test for archive_cleanup_command and recovery_end_comman Michael Paquier <[email protected]>
2021-12-17 07:00 [PATCH v9 5/5] Move code to apply one WAL record to a subroutine. Heikki Linnakangas <[email protected]>
2024-11-28 15:33 Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row Kirill Reshke <[email protected]>
2024-12-13 03:15 ` Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row jian he <[email protected]>
2025-01-08 07:05 ` Re: Change COPY ... ON_ERROR ignore to ON_ERROR ignore_row Kirill Reshke <[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