On 22/08/12 16:53, Ludovic Marcotte wrote:
On 22/08/12 15:47, Donny Brooks wrote:
For the past two days at random times we have been having this
happen. When it happens the sogod process eats 100% of the cpu and
just becomes unresponsive. To the point that even after stopping,
killing the stuck sogod processes, and restarting the service has it
hanging again within a minute of starting it back up. When this
happens it does not matter if we shut the machine down, restart it,
stop the process, etc when it comes back up it will happen within a
minute of starting. That is until it magically decides to run right.
Our sogo installation uses external imap/mysql/ldap servers and when
this has been happening all of the external servers are fine. Below
is an excerpt of the log. Any ideas?
That happens because SOGo is waiting for something from an external
component (LDAP, SQL, IMAP or SMTP). Use gdb and produce a stack
trace, that will tell you what it's waiting for.
There is any possiblity to add a functionality to the logging system for
log the process or connection that makes that SOGo eats all the cpu and
becomes unresponsive?
I think that such functionality will be very useful to debug problems
with external's SOGo components.
Thanks
--
[email protected]
https://inverse.ca/sogo/lists
