I notice in the log entries for the out of memory events have no username, database name or host
identifier (while regular logged events do) Does that mean anything to anybody?
Aug 28 08:25:50 db-1 postgres[29019]: [ID 748848 local0.warning] [111900-1] 2007-08-28
08:25:50.081 CDT 29019ERROR: out of memory
Aug 28 08:25:50 db-1 postgres[29019]: [ID 748848 local0.warning] [111900-2] 2007-08-28
08:25:50.081 CDT 29019DETAIL: Failed on request of size 536870910.
(regular log entry)
Aug 28 08:26:45 db-1 postgres[28785]: [ID 748848 local0.info] [114999-1] 2007-08-28 08:26:45.413
CDT jboss prod 192.168.20.44 28785LOG: duration: 22606.146 ms execute <unnamed>: select
--- Tom Lane <tgl@xxxxxxxxxxxxx> wrote:
> Can you correlate these occurrences with anything in the regular system
> logs (kernel log in particular)? The Postgres log shows nothing out of
> the ordinary --- it's simply that the kernel won't give us 512M for some
> reason. I'm guessing it's got something to do with overall system load.
>
> regards, tom lane
>
____________________________________________________________________________________
Moody friends. Drama queens. Your life? Nope! - their life, your story. Play Sims Stories at Yahoo! Games.
http://sims.yahoo.com/
---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings
