Amavisd going crazy - best way to restart it?

[Yann2]

Hello,

it seems that the antivirus just went crazy at 1pm today - it is using 100% of the cpu for nearly 3 hours now. What is the best way to restart it? I am not sure about just killing the process...

[jholder]

Hi Yann-
I'm sorry you're having difficulty. Have you tried restarting the clamd process, or bringing Zimbra down and restarting it?

There is a known issue with clamav in 4.5.6, and you can find out more about the work around here:
[SOLVED] Clamav problem ? What's happening ?

Let me know if this helps. . if not, we can try some different stuff.

Best,
john

[Yann2]

I did :
zimbra@mail:/home/yann$ zmamavisdctl restart

But it didn't seem to help... I found this additional error:

zimbra@mail:/home/yann$ zmantivirusctl restart
/opt/zimbra/bin/zmantivirusctl start|stop|reload|status
zimbra@mail:/home/yann$ zmantivirusctl reload
Usage: /opt/zimbra/bin/zmmtaconfigctl start|stop|kill|restart|status
Usage: /opt/zimbra/bin/zmamavisdctl start|stop|kill|restart|status
Usage: /opt/zimbra/bin/zmclamdctl start|stop|kill|restart|status


Then as root I tried a /etc/init.d/zimbra restarted which froze the complete system... I then logged in as zimbra and zmcontrol stop and startd, and it is up again, but I got 3 minutes of unwarned downtime /o\

Concerning the post you refered before, I am running 4.5.6.1044, the timeout setting is on 120 already - should i put it on 180?

[Edit] Btw, it is not the first time I encounter that reload/restart confusion in the CLI interface, I think other tools are also affected. I will check into bugzilla if there is already a bug filed for that

[jholder]

Originally Posted by Yann2

I did :
zimbra@mail:/home/yann$ zmamavisdctl restart

But it didn't seem to help... I found this additional error:

zimbra@mail:/home/yann$ zmantivirusctl restart
/opt/zimbra/bin/zmantivirusctl start|stop|reload|status
zimbra@mail:/home/yann$ zmantivirusctl reload
Usage: /opt/zimbra/bin/zmmtaconfigctl start|stop|kill|restart|status
Usage: /opt/zimbra/bin/zmamavisdctl start|stop|kill|restart|status
Usage: /opt/zimbra/bin/zmclamdctl start|stop|kill|restart|status


Then as root I tried a /etc/init.d/zimbra restarted which froze the complete system... I then logged in as zimbra and zmcontrol stop and startd, and it is up again, but I got 3 minutes of unwarned downtime /o\

Concerning the post you refered before, I am running 4.5.6.1044, the timeout setting is on 120 already - should i put it on 180?

[Edit] Btw, it is not the first time I encounter that reload/restart confusion in the CLI interface, I think other tools are also affected. I will check into bugzilla if there is already a bug filed for that

Yeah-
I'd increase it to 240.

So are you still getting defferred mail?

[Yann2]

Just changed that value to 240. Which deferred messages are you speaking from? I had a problem with amavis using 100%^of the cpu

[mmorse]

Usually the cpu thing is related to the startup of clamav, it's a good idea to check your queue's because if you had any mail at that time it'll be in the deferred queue. Yes they'll eventually deliver on their own-but you might want to send them through now instead of waiting.