Re: [Exim] How do you handle dead clams?

Top Page
Delete this message
Reply to this message
Author: hank
Date:  
To: exim-users
Subject: Re: [Exim] How do you handle dead clams?
On Tue, Jan 27, 2004 at 09:02:53AM +0000, Tim Jackson wrote:
> Hi Elliot, on Mon, 26 Jan 2004 11:11:46 -0700 you wrote:
>
> > clamd dies every so often which causes my users to get a 451.
>
> This is not an "exim-domain" solution, but you could run some kind of
> monitoring software such as monit ( http://www.tildeslash.com/monit/ )
> which can restart clamd automatically if it dies.


I have noticed a similar problem in which the clamd process doesn't
actually die, but it incorrectly accepts client requests, delivering
the same confusing error to my users. Unless I am misinformed, even
daemontools would not know to restart the server in this situation.
A default policy of "always let mail through, even if it has a virus"
is absolutely needed in order for mail to be scanned for viruses on
my network.

Clamd is a exceptional program, and I am very excited to help iron out
the wrinkles. While doing so, however, I cannot impact the delivery
of email. daemontools and monit are very useful as well, but
if there is a chance of handing the same SMTP error to users while the
process restarts, even for a c-hair of a second, then I am very hesitant.

Sorry for such a long post which is borderline offtopic. I have yet
to subscribe to the exiscan list, and acknowledge this discussion is
best sent there. I even admit to not have searched deeply into the
list archives. I just want it known that this may be a "exiscan-domain"
problem, which may be what you implied.

hank