Re: Backend crash with user defined aggregate - Mailing list pgsql-general

From Ron Mayer
Subject Re: Backend crash with user defined aggregate
Date
Msg-id 4335BEC8.8090606@cheapcomplexdevices.com
Whole thread Raw
In response to Re: Backend crash with user defined aggregate  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
Tom Lane wrote:
>
> Hmm.  Now that I look at it again, that message implies that something
> sent SIGTERM to the backend.  You need to figure out where that signal
> is coming from and why.  ...
> is it possible that you've started your postmaster with finite limits
> on process runtime?  What is the platform exactly, anyway?

Thanks. That's something for me to look into monday.  I don't think
there was a process runtime, but there might have been a memory limit.
The platform was a SuSE 9.1 on x86 but with so many updates & upgrades
from so many sources (rpm, apt, rcd, yast, tar.gz) that it's hard to
tell what it really is anymore.   Monday I'll try to reproduce on a
cleaner system and/or look for the SIGTERM with strace or something.

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Slow connection to the database
Next
From: "Jim C. Nasby"
Date:
Subject: Re: How many insert + update should one transaction