Commit214fa27

committed

Initialize shared memory copy of ckptXidEpoch correctly when not in recovery.

This bug was introduced by commit20d98ab,so backpatch this to 9.0-9.2 like that one.This fixes bug #6710, reported by Tarvi Pillessaar

1 parent90c85e4 commit214fa27Copy full SHA for 214fa27

File tree

-4

lines changed

-4

lines changed

Lines changed: 3 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -6249,11 +6249,14 @@ StartupXLOG(void)`
`6249`	`6249`	`ereport(PANIC,`
`6250`	`6250`	`(errmsg("invalid next transaction ID")));`
`6251`	`6251`
	`6252`	`+/* initialize shared memory variables from the checkpoint record */`
`6252`	`6253`	`ShmemVariableCache->nextXid=checkPoint.nextXid;`
`6253`	`6254`	`ShmemVariableCache->nextOid=checkPoint.nextOid;`
`6254`	`6255`	`ShmemVariableCache->oidCount=0;`
`6255`	`6256`	`MultiXactSetNextMXact(checkPoint.nextMulti,checkPoint.nextMultiOffset);`
`6256`	`6257`	`SetTransactionIdLimit(checkPoint.oldestXid,checkPoint.oldestXidDB);`
	`6258`	`+XLogCtl->ckptXidEpoch=checkPoint.nextXidEpoch;`
	`6259`	`+XLogCtl->ckptXid=checkPoint.nextXid;`
`6257`	`6260`
`6258`	`6261`	`/*`
`6259`	`6262`	`* We must replay WAL entries using the same TimeLineID they were created`
`@@ -6352,10 +6355,6 @@ StartupXLOG(void)`
`6352`	`6355`	`/* No need to hold ControlFileLock yet, we aren't up far enough */`
`6353`	`6356`	`UpdateControlFile();`
`6354`	`6357`
`6355`		`-/* initialize shared-memory copy of latest checkpoint XID/epoch */`
`6356`		`-XLogCtl->ckptXidEpoch=ControlFile->checkPointCopy.nextXidEpoch;`
`6357`		`-XLogCtl->ckptXid=ControlFile->checkPointCopy.nextXid;`
`6358`		`-`
`6359`	`6358`	`/* initialize our local copy of minRecoveryPoint */`
`6360`	`6359`	`minRecoveryPoint=ControlFile->minRecoveryPoint;`
`6361`	`6360`

Comments

(0)