Beth,
Can you rule out this bug as a cause?
https://bugs.launchpad.net/evergreen/+bug/1764542 Pretty sure it doesn't
explain slowness, but if you haven't applied the fix for that bug, you'll
likely see segmentation fault errors in the PostgreSQL logs and no results
returned from searches.
Yes. That was copied from the config. Ours was 1/4 of that at 5mil which
was causing issues.
Equinox suggested the change as it would crash our OpenSRF services every
once in a while
On Thu, 20 Dec 2018 at 11:56, Beth Longwell wrote:
> Chris,
>
> Thanks for the suggestions. I checked and our
Chris,
Thanks for the suggestions. I checked and our max_stanza_size is set to 2
million or 200. Can you confirm that your change was to 20 million or
2000 as your email specified?
As a part of the upgrade all our records were reingested so we should be
good on that front.
Beth
On Wed,
I ran across an issue like this just recently.
The ejabberd settings around line 125 with max_stanza_size was too small.
Equinox helped me out and we quadrupled it to 2000.
This does require a restart of ejabberd, opensrf and apache though.
Ejabberd is the XMPP server that communicates