Thanks for the response Magnus. The server logs do not show anything unusual, other than the occurrence 5 minutes prior. I will still take your advice on the upgrade.
2007-02-15 19:54:13 LOG: autovacuum: processing database "MM" 2007-02-15 19:55:13 LOG: autovacuum: processing database "postgres" 2007-02-15 19:55:37 ERROR: could not read block 558 of relation 1663/16403/16599: Invalid argument 2007-02-15 19:55:37 ERROR: current transaction is aborted, commands ignored until end of transaction block 2007-02-15 19:56:38 LOG: autovacuum: processing database "MM" 2007-02-15 19:57:38 LOG: autovacuum: processing database "postgres" 2007-02-15 19:58:38 LOG: autovacuum: processing database "MM" 2007-02-15 19:59:38 LOG: autovacuum: processing database "postgres" 2007-02-15 20:00:38 LOG: autovacuum: processing database "MM" 2007-02-15 20:01:38 LOG: autovacuum: processing database "postgres" 2007-02-15 20:02:38 LOG: autovacuum: processing database "MM" 2007-02-15 20:03:38 LOG: autovacuum: processing database "postgres" 2007-02-15 20:04:38 LOG: autovacuum: processing database "MM" ..... > -----Original Message----- > From: Magnus Hagander [mailto:[EMAIL PROTECTED] > Sent: Friday, February 23, 2007 3:29 AM > To: Craig White > Cc: pgsql-bugs@postgresql.org > Subject: Re: [BUGS] BUG #3032: Commit hung for days > > Craig White wrote: > > The following bug has been logged online: > > > > Bug reference: 3032 > > Logged by: Craig White > > Email address: [EMAIL PROTECTED] > > PostgreSQL version: 8.1.5 > > Operating system: Windows XP > > Description: Commit hung for days > > Details: > > > > I'm not looking for quick answers or a fix, but I wanted to submit this > for > > your information. > > > > Setup: > > > > Java App using Hibernate 3.2.1, C3P0 connection pool, JDBC to PostgreSQL > > database. > > > > PostgreSQL setup is pretty much the defaults. > > > > Occurrence: > > > > During some load testing of my application (best characterized as a > > multi-threaded transaction processing system), I ran into a transaction > that > > appeared to get stuck in its Commit. Using pgAdminIII, the 'Server > Status' > > window shows the Commit Query from 4 days prior. Other connections > > periodically execute a statement quickly and successfully so there are > not 2 > > statements in deadlock. > > > > The commit would have been altering a small number of rows from a couple > > tables. > > > > The commit happened at 2007/02/15 20:02:10, the only odd occurrence in > my > > logs occurred 5 minutes earlier. I'm not sure if it is releated or not. > > There was a lot of successful activity between 19:57 and 20:02. I'm not > yet > > sure of the cause of this error, but the prior error log is as follows: > > The interesting thing would be to see the server logs, not the > application logs. Specifically, an issue that could look just like this > was fixed in 8.1.7, in which case you would see weird error messages > about permission denied or such in the *server* logs. None of that would > show up in the client logs. > > So I would suggest upgrading to the latest release in 8.1.x. > > //Magnus ---------------------------(end of broadcast)--------------------------- TIP 6: explain analyze is your friend