Qmail keeps stopping

[sieb@]

Just upgraded one server 7.5.4 and after turning on Watchdog, I started getting notices every few hours saying Qmail stopped responding, then 10 mins later it would say it was fine. I tried to connect to qmail while it was down and it was indeed not responding (though submission responded), so I started Qmail and it came up fine. All the logs say is that Qmail stopped, then restarted, no errors.

Anyone else see this?

 

[d003]

Same problems here... I have now disabled the watchdog to see if that is causing the problem.

 

[b0nj]

i have the same issue with qmail, except it started before i upgraded to 7.5.4

any ideas for cure yet?

cheers

 

[Whistler]

As posted earlier - I had the same problem.

I've ended up disabled watchdog - as it seems to stop the running Qmail Daemon - I haven't had time to dig more into the problem yet, but hope to find some time over the weekend.

But please keep us posted if anyone finds the solution

 

[sieb@]

Hmm.. Qmail hasn't been a problem today. I will try disabling watchdog also and see what happens..

 

[b0nj]

ive gotta say that qmail is behaving itself today too and i still have watchdog running...hmmmm very strange

 

[Pagemakers]

Same issue here. Watchdog was sending me these emails:

[Sun, 28 Aug 2005 01:01:19 +0100] qmail on p15167366.pureserver.info: Connection failed.
'qmail' failed protocol test [SMTP] at INET[localhost:25].

Then:

[Sun, 28 Aug 2005 01:11:25 +0100] qmail on p15167366.pureserver.info: Connection passed.
'qmail' connection passed

I have disabled the Qmail monitoring but left Watchdog running.

I have never had a problem before installing Watchdog.

I have written to Plesk Tech Support and will keep you updated when they reply to me.

 

[Pagemakers]

Plesk tell me they know about this issue and are working on a solution.

 

[jamesyeeoc]

Did they confirm that it is a problem between Qmail and Watchdog?

 

[Pagemakers]

Yes

 

[Whistler]

I've also disovered today that DrWeb seems to have stoped working with the Control Panel Settings - did some statistics and discovered that no mailboxes (except a couple of my own) was checked - even though I've enabled virusscan in the controlpanel.

Had to manually enable scanning for all incoming mail until this problem is resolved.

Arrghhh

 

[jamesyeeoc]

Wow, and we even had to wait an extra month or so for this release. I'm glad I don't feel the urge to be on the 'bleeding edge'. Although I had planned on doing the updater this weekend on a clone of one of my production servers. Just haven't had the time yet..

 

[Whistler]

I actually mostly sadly most say - I've upgraded from 7.1.7 to 7.5.3 without having enough time to check which things were broken.

This and the fact I had to upgrade Perl from the old old old 5.005 version to the recent 5.8.7 got me off the track - not knowing what was broken by the Plesk upgrade and what went broken by the Perl upgrade.

I was then told the 7.5.4 release would fix several of my problems so I had to do the upgrade again. Now leaving me in the empty space of not knowing where things stopped working - but hopefully I'll manage to get back on track over the next weeks

But well, it keeps me busy :-D

 

[jamesyeeoc]

Yes, I seem to recall some of your other posts. Sorry that you have had such troubles. Would it not be faster to do a fresh install of Plesk on a new drive rather than taking 'weeks' to fix the current malfunctioning install?

 

[Whistler]

Well, you've only learn from mistakes and errors - so I'm putting up with it. Learning more and more

But I've considered the possibility I must admit But as said before - Sw-Support has done a great deal of helping me, providing suggestions on how to restore settings from database to the systemfiles etc. - so I think I'm on top of it - yet

 

[jamesyeeoc]

I agree about 'learning from mistakes', but I'd rather learn on a non-production server Whoops, I made an assumption that your problem(s) are on a live server. Best of luck to you!

 

[web-vision]

Same problem here:
[Sun, 28 Aug 2005 10:36:22 +0200] qmail on www3.web-vision.de: Connection passed.
'qmail' connection passed

[Mon, 29 Aug 2005 04:33:36 +0200] qmail on www3.web-vision.de: Connection passed.
'qmail' connection passed

What does that mean "passed"?

I guess I will deactivate watchdog now.

Regards,

Boris

 

[jamesyeeoc]

What does that mean "passed"?

That means that Watchdog was able to establish a connection to the qmail service, so it 'passed' the test.

A failure would look like what Pagemakers posted earlier:

[Sun, 28 Aug 2005 01:01:19 +0100] qmail on p15167366.pureserver.info: Connection failed.
'qmail' failed protocol test [SMTP] at INET[localhost:25].

 

[Pagemakers]

Also woke up to this, this morning:


[Mon, 29 Aug 2005 04:05:12 +0100] courier-imap on p15167477.pureserver.info: Connection failed.
'courier-imap' failed protocol test [IMAP] at INET[localhost:143].

[Mon, 29 Aug 2005 04:15:42 +0100] courier-imap on p15167477.pureserver.info: Connection passed.
'courier-imap' connection passed.

IMAP as well. Although it does restart itself.

I wonder if this was happening because the domain was backup up at the time, or if if 4PSA's Total Backup was running (which it does at 4am each night).

 

[jamesyeeoc]

One easy way to find out - turn off all backups for a single night and see if it happens again. I don't like that suggestion myself, but I can't think of anything else which would be as simple a test.

Of course, the time could just be coincidence, your earlier posted failures were around 1 am