rdelaage/postgresql
1
0
Fork 0
mirror of https://git.postgresql.org/git/postgresql.git synced 2024-09-27 22:22:05 +02:00
Code Issues Packages Projects Releases Wiki Activity

Fix WAL-logging of setting the visibility map bit.

Browse Source 
The operation that removes the remaining dead tuples from the page must
be WAL-logged before the setting of the VM bit. Otherwise, if you replay
the WAL to between those two records, you end up with the VM bit set, but
the dead tuples are still there.

Backpatch to 9.3, where this bug was introduced.
This commit is contained in:
Heikki Linnakangas 2013-12-13 13:52:47 +02:00
parent 46328916ee
commit a49633d8dc
1 changed files with 12 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
src/backend/commands/vacuumlazy.c
View File

@ -1182,12 +1182,21 @@ lazy_vacuum_page(Relation onerel, BlockNumber blkno, Buffer buffer,
/*
* Mark buffer dirty before we write WAL.
*
* If checksums are enabled, visibilitymap_set() may log the heap page, so
* we must mark heap buffer dirty before calling visibilitymap_set().
*/
MarkBufferDirty(buffer);
/* XLOG stuff */
if (RelationNeedsWAL(onerel))
{
XLogRecPtr recptr;
recptr = log_heap_clean(onerel, buffer,
NULL, 0, NULL, 0,
unused, uncnt,
vacrelstats->latestRemovedXid);
PageSetLSN(page, recptr);
}
/*
* Now that we have removed the dead tuples from the page, once again
* check if the page has become all-visible.
@ -1201,18 +1210,6 @@ lazy_vacuum_page(Relation onerel, BlockNumber blkno, Buffer buffer,
visibility_cutoff_xid);
}
/* XLOG stuff */
if (RelationNeedsWAL(onerel))
{
XLogRecPtr recptr;
recptr = log_heap_clean(onerel, buffer,
NULL, 0, NULL, 0,
unused, uncnt,
vacrelstats->latestRemovedXid);
PageSetLSN(page, recptr);
}
END_CRIT_SECTION();
return tupindex;