On 6/9/2017 1:33 AM, Paul R. Ganci wrote: > For well over a year and half I have had two incoming mail servers > running postfix + amavisd + spamassassin + clamd on a CentOS 7 > system working flawlessly. Over the past two days something has > happened such that the postfix stops delivering messages to user > inboxes with this message: > > Jun 8 19:37:05 mx01 postfix/error[22985]: 7B867836338F: > to=<xx...@xxxxxx.net>, relay=none, delay=295, > delays=0.49/294/0/0.05, dsn=4.4.2, status=deferred (delivery > temporarily suspended: conversation with 127.0.0.1[127.0.0.1] timed > out while receiving the initial server greeting) > > For whatever reason it appears that the clamd stops working and the > backup clamscan does not takeover. As a result postfix stops mail > delivery and begins to queue messages. When I restart amavisd and > then flush the postfix queue the messages begin to be delivered > again. Then for several hours everything works normally until some > event causes clamd to stop again and the messages begin to queue again. > > I have no idea where to begin to debug this problem. For the time > being I created a cron job to fix the problem. But I definitely need > to find the root cause of the issue. What bugs me as that everything > has worked flawlessly for a year and a half and all of the sudden I > have begun to have problems. Other than the postfix error message I > don't see anything in the logs. Can anybody suggest a good way to > debug the problem? > > Also, has anybody else seen such a problem? If so, what did you do > to fix it? >
First, this isn't a postfix problem. Postfix is reporting a problem downstream. Whatever fix is needed will be outside of postfix. We use monit to alert the admin and restart clamd when it gets stuck, which isn't very often. IMO using clamscan as a backup scanner is so slow as to be useless unless you have a very light mail flow. Another choice is to enable the "always pass" backup antivirus stub in the amavis config so that if clamd dies mail will continue flowing, but unscanned. This may or may not be related, but 2 days ago the clam team released a huge signature update that slower machines are having trouble applying. Check to see if your freshclam is stuck or failing. Followups should go to either the clamav or amavisd lists, but you'll need to provide more information to get a better answer. -- Noel Jones
