From: | "Joey Adams" <jea(at)tcdi(dot)com> |
---|---|
To: | <pgsql-general(at)postgresql(dot)org> |
Subject: | FlushRelationBuffers Error |
Date: | 2006-10-02 14:48:37 |
Message-ID: | 313F815E37F18C48B4DCF6D9E77BCC8A0B9B74@tcdiexch2k3.tcdi.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Last night during a vacuum full, one of our production 7.4.12 databases threw this warning:
WARNING: FlushRelationBuffers("idxtype26", 6137): block 5752 is referenced (private 0, global 1)
vacuumdb: vacuuming of database "cvdb" failed: ERROR: FlushRelationBuffers returned -2
PANIC: cannot abort transaction 35650725, it was already committed
server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
We do use some external C functions that deal with creating indexes and locking tables, but we're pretty confident that they're locking correctly. Another bit of information, the database gets restarted before the vacuum process begins.
Is this something that will resolve itself the next time the database gets vacuumed or does it need some kind of immediate action? Does the index need reindexed?
Thanks
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2006-10-02 14:59:42 | Re: Major Performance decrease after some hours |
Previous Message | Ardian Xharra | 2006-10-02 13:38:30 | Re: How to trace index to table? |