Hi Ivar,
I agree with Magnus. As addition, also i would change:
effective_cache_size
and
maintenance_work_mem
Then disable OOM killer, change logging in order to log files on disk to see if
your work_mem is too low and spills on disk, rethink your data structure and,
overall, I think that
On Tue, Jun 26, 2018 at 11:40 AM, Ivar Fredriksen wrote:
> A single large query is able to spend all the system memory (as reported
> by top), and the oom_reaper kills postgres. See bottom of email for an
> example query and logs.
>
>
>
> Expected behavior would be that postgres is not killed and
