On 2011 Feb 11, at 13:54 , Jan-Frode Myklebust wrote: > For us it took down clamd on 15 servers at 00:03 today, and > we received the fix 3 hours later... but clamd wasn't restarter > before later this morning, leading to huge mailqueues. > > We should probably look into verifying the db before telling > clamd to reload it...
I suggest you instead look at your mail config, verifying that mail keeps on flowing when clamav happens to be down/unresponsive. Unless you want to err on the safe side, and have a policy in place that says "we do not want to receive/send ANY mail when the virus scanning doesn't work". In that case, your system is already working as designed, and tonight's outage was actually helpful, because it prevented mail from getting through that could have been detected by a newer version of the database. On the other hand, since you haven't updated ClamAV in over a year, leading to (significantly) decreased detection, maybe the scanning of email isn't top priority, and your mail scanning engine needs to fallback to letting mail through on scan errors. -- Jan-Pieter Cornet <joh...@xs4all.nl> "People are continuously reinventing the flat tyre". _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://www.clamav.net/support/ml
