Craig McElroy <craig.mcelroy@contegix.com> writes:
> Core was generated by `/usr/local/pgsql/bin/postgres -D /pgdata01/data'.
> Program terminated with signal 11, Segmentation fault.
> #0 0x080b8ee0 in entrySplitPage ()
> #1 0x080baccf in ginInsertValue ()
> #2 0x080b81b7 in gin_xlog_cleanup ()
> #3 0x080af4ce in StartupXLOG ()
> #4 0x080c04ca in BootstrapMain ()
> #5 0x08186b2f in StartChildProcess ()
> #6 0x081889eb in PostmasterMain ()
> #7 0x0814ee9e in main ()
Hm, I wonder if this is explained by a bug already fixed in 8.2.5:
2007-06-04 11:59 teodor
* src/backend/access/gin/: gindatapage.c, ginentrypage.c, ginget.c,
ginvacuum.c, ginxlog.c (REL8_2_STABLE): Fix bundle bugs of GIN: -
Fix possible deadlock between UPDATE and VACUUM queries. Bug never
was observed in 8.2, but it still exist there. HEAD is more
sensitive to bug after recent "ring" of buffer improvements. -
Fix WAL creation: if parent page is stored as is after split then
incomplete split isn't removed during replay. This happens rather
rare, only on large tables with a lot of updates/inserts. - Fix
WAL replay: there was wrong test of XLR_BKP_BLOCK_* for left page
after deletion of page. That causes wrong rightlink field: it
pointed to deleted page. - add checking of match of clearing
incomplete split - cleanup incomplete split list after proceeding
All of this chages doesn't change on-disk storage, so backpatch...
But second point may be an issue for replaying logs from previous
version.
Teodor, can you comment on whether this stack trace looks like it could
be related to that fix?
Craig, can you retry your test scenario on 8.2.5?
regards, tom lane