Fix MVCC bug with prepared xact with subxacts on standby

We did not recover the subtransaction IDs of prepared transactions
when starting a hot standby from a shutdown checkpoint. As a result,
such subtransactions were considered as aborted, rather than
in-progress. That would lead to hint bits being set incorrectly, and
the subtransactions suddenly becoming visible to old snapshots when
the prepared transaction was committed.

To fix, update pg_subtrans with prepared transactions's subxids when
starting hot standby from a shutdown checkpoint. The snapshots taken
from that state need to be marked as "suboverflowed", so that we also
check the pg_subtrans.

Backport to all supported versions.

Discussion: https://www.postgresql.org/message-id/6b852e98-2d49-4ca1-9e95-db419a2696e0@iki.fi

2 files changed, 18 insertions, 6 deletions

diff --git a/src/backend/storage/ipc/procarray.c b/src/backend/storage/ipc/procarray.c
index d5165aa0d9f..387b4a405b0 100644
--- a/src/backend/storage/ipc/procarray.c
+++ b/src/backend/storage/ipc/procarray.c
@@ -1106,7 +1106,7 @@ ProcArrayApplyRecoveryInfo(RunningTransactions running)
 		 * If the snapshot isn't overflowed or if its empty we can reset our
 		 * pending state and use this snapshot instead.
 		 */
-		if (!running->subxid_overflow || running->xcnt == 0)
+		if (running->subxid_status != SUBXIDS_MISSING || running->xcnt == 0)
 		{
 			/*
 			 * If we have already collected known assigned xids, we need to
@@ -1258,7 +1258,7 @@ ProcArrayApplyRecoveryInfo(RunningTransactions running)
 	 * missing, so conservatively assume the last one is latestObservedXid.
 	 * ----------
 	 */
-	if (running->subxid_overflow)
+	if (running->subxid_status == SUBXIDS_MISSING)
 	{
 		standbyState = STANDBY_SNAPSHOT_PENDING;
 
@@ -1270,6 +1270,18 @@ ProcArrayApplyRecoveryInfo(RunningTransactions running)
 		standbyState = STANDBY_SNAPSHOT_READY;
 
 		standbySnapshotPendingXmin = InvalidTransactionId;
+
+		/*
+		 * If the 'xids' array didn't include all subtransactions, we have to
+		 * mark any snapshots taken as overflowed.
+		 */
+		if (running->subxid_status == SUBXIDS_IN_SUBTRANS)
+			procArray->lastOverflowedXid = latestObservedXid;
+		else
+		{
+			Assert(running->subxid_status == SUBXIDS_IN_ARRAY);
+			procArray->lastOverflowedXid = InvalidTransactionId;
+		}
 	}
 
 	/*
@@ -2833,7 +2845,7 @@ GetRunningTransactionData(void)
 
 	CurrentRunningXacts->xcnt = count - subcount;
 	CurrentRunningXacts->subxcnt = subcount;
-	CurrentRunningXacts->subxid_overflow = suboverflowed;
+	CurrentRunningXacts->subxid_status = suboverflowed ? SUBXIDS_IN_SUBTRANS : SUBXIDS_IN_ARRAY;
 	CurrentRunningXacts->nextXid = XidFromFullTransactionId(TransamVariables->nextXid);
 	CurrentRunningXacts->oldestRunningXid = oldestRunningXid;
 	CurrentRunningXacts->oldestDatabaseRunningXid = oldestDatabaseRunningXid;
diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
index 87b04e51b36..872679ca447 100644
--- a/src/backend/storage/ipc/standby.c
+++ b/src/backend/storage/ipc/standby.c
@@ -1184,7 +1184,7 @@ standby_redo(XLogReaderState *record)
 
 		running.xcnt = xlrec->xcnt;
 		running.subxcnt = xlrec->subxcnt;
-		running.subxid_overflow = xlrec->subxid_overflow;
+		running.subxid_status = xlrec->subxid_overflow ? SUBXIDS_MISSING : SUBXIDS_IN_ARRAY;
 		running.nextXid = xlrec->nextXid;
 		running.latestCompletedXid = xlrec->latestCompletedXid;
 		running.oldestRunningXid = xlrec->oldestRunningXid;
@@ -1349,7 +1349,7 @@ LogCurrentRunningXacts(RunningTransactions CurrRunningXacts)
 
 	xlrec.xcnt = CurrRunningXacts->xcnt;
 	xlrec.subxcnt = CurrRunningXacts->subxcnt;
-	xlrec.subxid_overflow = CurrRunningXacts->subxid_overflow;
+	xlrec.subxid_overflow = (CurrRunningXacts->subxid_status != SUBXIDS_IN_ARRAY);
 	xlrec.nextXid = CurrRunningXacts->nextXid;
 	xlrec.oldestRunningXid = CurrRunningXacts->oldestRunningXid;
 	xlrec.latestCompletedXid = CurrRunningXacts->latestCompletedXid;
@@ -1366,7 +1366,7 @@ LogCurrentRunningXacts(RunningTransactions CurrRunningXacts)
 
 	recptr = XLogInsert(RM_STANDBY_ID, XLOG_RUNNING_XACTS);
 
-	if (CurrRunningXacts->subxid_overflow)
+	if (xlrec.subxid_overflow)
 		elog(DEBUG2,
 			 "snapshot of %d running transactions overflowed (lsn %X/%X oldest xid %u latest complete %u next xid %u)",
 			 CurrRunningXacts->xcnt,