On 9/11/2014 3:12 PM, Marcel Waldvogel wrote:
> How many users do you have? How many SOGo processes do you allow? Could
> it be that there are timeouts or connection aborts somewhere in the
> chain (client→network→Web server→SOGo→database)? Do any of the log files
> indicate anything along these lines?
> 

We only have half a dozen users testing out SOGo (because of the
problems, I can't roll it out to additional people).

PostgreSQL is configured to allow 1000 connections.  There are no errors
in the pg_log when this happens.

Client -> network -> Web Server is over a T1 (1.5Mbps) line, which goes
down maybe once per year (very reliable).  And different clients see the
prompt at different times of the day (there is no pattern).

Apache HTTPD / SOGo / PostgreSQL are all on the same machine.

SELinux is running in permissive mode until I get it working well enough
to use audit2allow to help me write a security policy.

Nothing in ssl_error_log or error_log for Apache HTTPD.

Not sure where SOGo logs errors.
-- 
[email protected]
https://inverse.ca/sogo/lists

Reply via email to