[Mimedefang] MD Lock-Ups occasionally

David F. Skoll dfs at roaringpenguin.com
Wed Oct 17 09:33:17 EDT 2007


Kevin A. McGrail wrote:

> Lately on just one machine, I am seeing MD just end up in a hung state.

That's strange.  Can you strace/ktrace/truss some of the processes to
see what's going on?

> 1 - What's the best command to run to see what might be occurring when
> it is locked up?

md-mx-ctrl slaves
md-mx-ctrl busyslaves

and then use OS-level tracing tools.

> 2 - What's the best command to run to simply check that MD is alive and
> if not restart the multiplexor and children?

You could have a script that checks if the pids in the MD and multiplexor
pid files are still running, I suppose.  We have a monitoring script for
CanIt, but it won't work with generic MIMEDefang because our CanIt
init script is very different.

> root     30588  0.0  0.1   7960  4760 ?        S    Oct16   0:15
> sendmail: ./l9DGFVLL020125 postoffice.telstra.net.: user open

Those look like network problems.  They are sendmail processes on
your end attempting to connect to remote hosts.

Regards,

David.



More information about the MIMEDefang mailing list