From: | Stephen Frost <sfrost(at)snowman(dot)net> |
---|---|
To: | Erik Rijkers <er(at)xs4all(dot)nl> |
Cc: | Christian Kruse <christian(at)2ndQuadrant(dot)com>, pgsql-hackers(at)postgresql(dot)org, teodor(at)sigaev(dot)ru, obartunov(at)gmail(dot)com, andrew(at)dunslane(dot)net |
Subject: | Re: nested hstore - large insert crashes server |
Date: | 2014-02-04 19:17:06 |
Message-ID: | 20140204191706.GU2921@tamriel.snowman.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
* Erik Rijkers (er(at)xs4all(dot)nl) wrote:
> On Tue, February 4, 2014 18:56, Christian Kruse wrote:
> > On 04/02/14 17:41, Erik Rijkers wrote:
> >> 2014-02-04 10:34:25.376 CET 29133 LOG: server process (PID 29459) was terminated by signal 9: Killed
> >
> > Did you check if this was the OOM killer? Should be logged in dmesg.
>
> I would be surprised if it wasn't. (no access to that machine at the moment)
>
> How do we regard such crashes? It seems to me this was rather eaasily 'provoked' (for want of a better word).
>
> I am inclined to blame the patch...
It sounds like there is at least some investigation which should happen
here to see why we're using so much memory (well beyond work_mem and
even maint_work_mem it sounds like), but it would also be good to have
the machine reconfigured to not allow OOM killing.
Thanks,
Stephen
From | Date | Subject | |
---|---|---|---|
Next Message | Adrian Klaver | 2014-02-04 19:18:14 | Re: narwhal and PGDLLIMPORT |
Previous Message | Andrew Dunstan | 2014-02-04 19:17:00 | Re: narwhal and PGDLLIMPORT |