IMAP service keeps failing and restarting

[consultorpc]

We are having the same problem with a server using Dovecot. Although we don't see any problem in the server at all it's annoying to get these errors mails every day.

imap failed @ Sun Apr 12 04:40:12 2009. A restart was attempted automagically.
Service Check Method:  [check command]

Cmd Service Check Raw Output: dovecot is not running

[Kent Brockman]

We are having the same problem with a server using Dovecot. Although we don't see any problem in the server at all it's annoying to get these errors mails every day.

Pretty strange.

Is the Dovecot daemon set to be monitored by Service Manager?
How many auth daemons are you running?
Do you have enough free memory? (Do the auth daemons use too much memory?)
Are you running heavy processes?
Do you have the latest STABLE cPanel version?

We increased the number of daemons to much more than 5. I think we used 40 or 50 and that does not appear to impact overall VPS performance.

Finally, we migrated to Dovecot and the login problems dissappeared. So, it must be some kind of problem with Courier.

[consultorpc]

Hi Kent,

Yes, the dovecot daemon is monitored by system manager.

We are using 5 Number of Spare Authentication Processes and 60 Maximum Number of Authentication Processes.

This is not a memory not load problem as even the error comes nightly when the server has very litte load.

And yes, we have the lastest cPanel and Dovecot versions running.

Never mind, I will try to increase these values to see if now the alerts disappear.

[kampret]

I have the same issue just right now, dovecot refused to run/start.
Force update dovecot, dovecot -n, everything seems like ignored by exim and pop and imap.
Still recognize as : Dovecote is not running.

Reverting to Courier makes everything to work now.
But I need to use dovecot again.

How to fix this dovecote issue?

Thank you.

[jandafields]

Current CPanel dovecot version is 1.1.11.

Current official dovecot version stable is 1.1.14.

Perhaps the newest stable version will fix issues???

[consultorpc]

Look if your date is the same as hwclock, we had issues with one server with this error.

[formula13x]

I am fighting this problem too.. But mine is much worse (about 40 times a day).
It started a few weeks ago. I am on a VPS. Any more things to try? Also, what were you talking about with the hwclock?

the authdaemon instances just keep dying.. Is there anything I can do to debug it to see why they are dying so quickly? I've upped them to 10.

[formula13x]

Here is a hack, just to keep it working I suppose.
I don't like doing this, and want a real solution not a hack.
But anyhow.
I just created a script named watchauthdaemond.sh
as

#!/bin/bash
if test `ps -elf|grep authdaemond|wc -l` -lt 2
then
authdaemond stop
authdaemond start
fi


Then I created a created an entry in my crontab to make it run every minute

* * * * * /scripts/watchauthdaemond.sh > /dev/null 2>&1

We'll see if it works for the time being.

Any real solution would be greatly appreciated.

[jols]

Hi, I've increased the number of auth daemons from 2 to 5 this week and there are still login problems. I've reinstalled exim and courier and the problem is still there. Any idea?

BTW, the imap problems reported originally in this post have disappeared.

Regarding, increasing "the number of auth daemons".
Pardon me, but can you list the file that needs to be changed, and/or the general process to do this? Thanks.

[jols]

Never mind. I found it:

WHM ---> Mailserver Configuration

[Kent Brockman]

Just for the record: since I upgraded to cPanel 11.24.4-S35075, the courier issue has disappeared. Great!