Re: [PERFORM] select on 22 GB table causes"An I/O error occured while sending to the backend." exception

>>> Bill Moran <wmoran@collaborativefusion.com> wrote:
> In response to Matthew Wakeling <matthew@flymine.org>:
>>
>> Probably the best solution is to just tell the kernel somehow to
never
>> kill the postmaster.
>
> This thread interested me enough to research this a bit.
>
> In linux, it's possible to tell the OOM killer never to consider
> certain processes for the axe, using /proc magic. See this page:
> http://linux-mm.org/OOM_Killer
>
> Perhaps this should be in the PostgreSQL docs somewhere?

That sure sounds like a good idea.

Even though the one time the OOM killer kicked in on one of our
servers, it killed a runaway backend and not the postmaster
( http://archives.postgresql.org/pgsql-bugs/2008-07/msg00105.php ),
I think I will modify our service scripts in /etc/init.d/ to pick off
the postmaster pid after a start and echo -16 (or some such) into the
/proc/<pid>/oom_adj file (which is where I found the file on my SuSE
system).

Thanks for the research and the link!

-Kevin

--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance