database/postgres
database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-05-17 06:41:24 +03:00
Code

Fix aboriginal mistake in lazy VACUUM's code for truncating away

Browse Source 
no-longer-needed pages at the end of a table.  We thought we could throw away
pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such
tuples very likely have index entries pointing at them, and we wouldn't have
removed the index entries.  The problem only emerges in a somewhat unlikely
race condition: the dead tuples have to have been inserted by a transaction
that later aborted, and this has to have happened between VACUUM's initial
scan of the page and then rechecking it for empty in count_nondeletable_pages.
But that timespan will include an index-cleaning pass, so it's not all that
hard to hit.  This seems to explain a couple of previously unsolved bug
reports.
This commit is contained in:
Tom Lane 2007-09-16 02:38:25 +00:00
parent 2f2f9b9bae
commit 11a8925afd
1 changed files with 12 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

68
src/backend/commands/vacuumlazy.c
View File

@ -31,7 +31,7 @@
* *
* *
* IDENTIFICATION * IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/commands/vacuumlazy.c,v 1.32.2.1 2005/05/07 21:33:21 tgl Exp $ * $Header: /cvsroot/pgsql/src/backend/commands/vacuumlazy.c,v 1.32.2.2 2007/09/16 02:38:25 tgl Exp $
* *
*------------------------------------------------------------------------- *-------------------------------------------------------------------------
*/ */
@ -798,7 +798,7 @@ lazy_truncate_heap(Relation onerel, LVRelStats *vacrelstats)
} }
/* /*
* Rescan end pages to verify that they are (still) empty of needed tuples. * Rescan end pages to verify that they are (still) empty of tuples.
* *
* Returns number of nondeletable pages (last nonempty page + 1). * Returns number of nondeletable pages (last nonempty page + 1).
*/ */
@ -806,7 +806,6 @@ static BlockNumber
count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats) count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats)
{ {
BlockNumber blkno; BlockNumber blkno;
HeapTupleData tuple;
/* Strange coding of loop control is needed because blkno is unsigned */ /* Strange coding of loop control is needed because blkno is unsigned */
blkno = vacrelstats->rel_pages; blkno = vacrelstats->rel_pages;
@ -816,9 +815,7 @@ count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats)
Page page; Page page;
OffsetNumber offnum, OffsetNumber offnum,
maxoff; maxoff;
bool pgchanged, bool hastup;
tupgone,
hastup;
CHECK_FOR_INTERRUPTS(); CHECK_FOR_INTERRUPTS();
@ -833,13 +830,12 @@ count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats)
if (PageIsNew(page) || PageIsEmpty(page)) if (PageIsNew(page) || PageIsEmpty(page))
{ {
/* PageIsNew robably shouldn't happen... */ /* PageIsNew probably shouldn't happen... */
LockBuffer(buf, BUFFER_LOCK_UNLOCK); LockBuffer(buf, BUFFER_LOCK_UNLOCK);
ReleaseBuffer(buf); ReleaseBuffer(buf);
continue; continue;
} }
pgchanged = false;
hastup = false; hastup = false;
maxoff = PageGetMaxOffsetNumber(page); maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber; for (offnum = FirstOffsetNumber;
@ -847,52 +843,16 @@ count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats)
offnum = OffsetNumberNext(offnum)) offnum = OffsetNumberNext(offnum))
{ {
ItemId itemid; ItemId itemid;
uint16 sv_infomask;
itemid = PageGetItemId(page, offnum); itemid = PageGetItemId(page, offnum);
if (!ItemIdIsUsed(itemid)) /*
continue; * Note: any non-unused item should be taken as a reason to keep
* this page. We formerly thought that DEAD tuples could be
tuple.t_datamcxt = NULL; * thrown away, but that's not so, because we'd not have cleaned
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid); * out their index entries.
tuple.t_len = ItemIdGetLength(itemid); */
ItemPointerSet(&(tuple.t_self), blkno, offnum); if (ItemIdIsUsed(itemid))
tupgone = false;
sv_infomask = tuple.t_data->t_infomask;
switch (HeapTupleSatisfiesVacuum(tuple.t_data, OldestXmin))
{
case HEAPTUPLE_DEAD:
tupgone = true; /* we can delete the tuple */
break;
case HEAPTUPLE_LIVE:
/* Shouldn't be necessary to re-freeze anything */
break;
case HEAPTUPLE_RECENTLY_DEAD:
/*
* If tuple is recently deleted then we must not
* remove it from relation.
*/
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
break;
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
/* check for hint-bit update by HeapTupleSatisfiesVacuum */
if (sv_infomask != tuple.t_data->t_infomask)
pgchanged = true;
if (!tupgone)
{ {
hastup = true; hastup = true;
break; /* can stop scanning */ break; /* can stop scanning */
@ -900,11 +860,7 @@ count_nondeletable_pages(Relation onerel, LVRelStats *vacrelstats)
} /* scan along page */ } /* scan along page */
LockBuffer(buf, BUFFER_LOCK_UNLOCK); LockBuffer(buf, BUFFER_LOCK_UNLOCK);
ReleaseBuffer(buf);
if (pgchanged)
WriteBuffer(buf);
else
ReleaseBuffer(buf);
/* Done scanning if we found a tuple here */ /* Done scanning if we found a tuple here */
if (hastup) if (hastup)