On 02/05/17 20:20, Kevin Bowen via Users wrote:
> Hello,
> I'm running policyd on a cluster of 4 postfix servers, talking to a
> dedicated mysql backend, and I'm seeing a fair amount of
> 'db_failure_bypass' in my logs - i.e. unable to contact the mysql
> server. Since I have policyd set to bypass_mode=pass, this isn't a
> huge deal, but I would like to resolve it if possible. The mysql
> server never seems to use more than single-digit percentages of its
> available cpu or ram, so that doesn't seem to be the issue. Can anyone
> point me to any relevant settings to tweak, either in policyd or
> mysql, that might help resolve the issue?


This happens when the MySQL API ping command gets a negative response.
I'd suggest checking the MySQL server logs, specifically for connection
count limits being exceeded.

_______________________________________________
Users mailing list
[email protected]
http://lists.policyd.org/mailman/listinfo/users_lists.policyd.org

Reply via email to