So I was forced to restart the SmarterMail service today due to the system being unresponsive as a result of running a Message Archive Query (not good folks, this needs fixing). I had inadvertantly entered 2/19/0019 as the starting date (and the system took it as gospel... again, not good). Anyway that made the system scan ALL Archive indexes on the machine and it ran and ran and ran. Which was SUPER slow too, far far slower than prior to updating to the 100.whatever the latest version is.
During that time the CPU was locked at about 90 to 95% and made the the spool backup, users unable to login, and even in the admin I could barely navigate around.
So... Made a judgement call and restarted the SM service.
That fixed the CPU issue, but introduced a new issue.
Following SM coming back online, the ClamAV stopped communicating with the requests from SM to process messages. Normally i have maybe 10 to 15 failed connections to ClamAV an hour... now there were THOUSANDS. And NO successful (aka New) connections (monitored through the charting). Looking in the ClamAV logs it was something about not loading the PUA files.
After fiddling with the AV processing settings and bumping up the timeout to 120 sec and the count to failure to 40 (saw that recommended here somewhere)... then stopping and restarting the "Enable ClamAV" setting in SM. After doing that several times, it finally seemed to take effect and start working again. However, I noticed that even when the setting in SM was "Enable ClamAV" OFF, I would still see the clamd.exe hopping around in the Process list of Task Manager.
So... the question...
If you MUST restart SM Service, what is the proper method and sequence of doing so so that you do not lose your communication with ClamAV? Or does the entire server need to be restarted instead of just the SM service for that to all sync up correctly?