From: | "Peter" <peter(at)greatnowhere(dot)com> |
---|---|
To: | |
Cc: | <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Transaction eating up all RAM |
Date: | 2006-03-13 17:43:25 |
Message-ID: | 01c501c646c5$a18ad610$a502a8c0@komtek.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general pgsql-hackers |
>> I have stored proc that retrieves a bunch of data, stores it in temp =
>> table, computes all sorts of totals/averages/whatnots from the temp =
>> table, and inserts results in another table. It works fine (except I =
>> don't like wrapping all SQL statements in 'execute'), but multiple calls
>> =
>> to that proc from another procedure causes excessive memory usage =
>> (upwards of 400M), and server eventually runs out of swap space. I =
>> believe this is because PG caches transactions in RAM, and this =
>> particular one is a bit too big.=20
>
> Your belief is incorrect.
>
> You could be looking at a memory-leak bug. Or, if there are foreign
> keys involving the tables, you could be looking at the list of pending
> foreign key trigger events getting too large. There's not enough
> information here to say.
I have no triggers defined on any of the tables, and no foreign keys that
could cause cascaded updates and stuff. Care to see full text of the proc?
It's pl/PgPerlU
Peter
From | Date | Subject | |
---|---|---|---|
Next Message | Michael Fuhr | 2006-03-13 17:50:02 | Re: Possible infinite loop in query using bitmap scans |
Previous Message | Andreas Kretschmer | 2006-03-13 16:46:50 | Re: dump from old version |
From | Date | Subject | |
---|---|---|---|
Next Message | Rodrigo Hjort | 2006-03-13 17:56:00 | Restoring a Full Cluster on a Different Architecture (32 x 64) |
Previous Message | Magnus Hagander | 2006-03-13 17:27:03 | Re: [PERFORM] Hanging queries on dual CPU windows |