[Mimedefang] Anyone else having problems with Clamd 0.65?

[Kelson Vibber]

After upgrading ClamAV to 0.65 a few weeks ago, we've had problems where 
MIMEDefang will simply stop closing down slaves.  Instead of timing out, MD 
slaves would just hang around forever, not pass anything through, and 
eventually max out and start tempfailing.  Often just restarting clamd 
would resolve the issue.

I'd removed clamd from our filter at one point, then added it back in 
yesterday in response to MyDoom/Novarg.  It wasn't long before the same 
thing happened again, so I removed it and now we're back to File::Scan and 
a few ad-hoc filter rules (thanks to those who posted them, by the way).

On one hand this sounds like a bug in clamd - but on the other, it seems to 
me that MIMEDefang should be timing out if it can't get anything back from 
clamd, and freeing up slaves instead of keeping them around.

(MD 2.39, RedHat 7.3*, Perl 5.6.1, ClamAV 0.65)

* We're watching the RH9 and other vulnerability lists, and using patches 
from Fedora Legacy.  Anything without a patch that looks like a problem is 
either removed or upgraded from source.  The plan is to replace the machine 
soon, at which point we can wipe this one and install something newer on 
this one, which is the main reason it hasn't been upgraded already.


Kelson Vibber
SpeedGate Communications <www.speed.net>