Thousands of postfix-queue processes

[jacko@]

Hello,

OS: CentOS release 5.4 (Final)
Plesk: psa-9.3.0-cos5.build93091230.06 (Plesk 9.3)
MTA: Postifx

# ps auxw | grep /usr/lib64/plesk-9.0/postfix-queue | wc -l
1158

1158 postfix-queue processes running.. this doesn't seem right?

....
30 32724 0.0 0.0 11520 840 ? Ss Feb18 0:00 /usr/lib64/plesk-9.0/postfix-queue 127.0.0.1 10027 before-queue
30 32725 0.0 0.0 11516 788 ? Ss Feb18 0:00 /usr/lib64/plesk-9.0/postfix-queue 127.0.0.1 10026 before-remote
....

Anyone else seeing this? This is basically a fresh install of Plesk 9.3 on Centos 5.4 (ART).

Thanks,
jacko

 

[IgorG]

This problem discussed in this thread - http://forum.parallels.com/showthread.php?p=405738

 

[jacko@]

Hello IgorG,

I'm not seeing any "queue file write error" messages in log files. Is this the same problem?

No errors at all in /usr/local/psa/var/log/maillog.

Thanks,
jacko

 

[IgorG]

If there are no any errors in maillog and mailserver works fine why these processes disturb you?

 

[jacko@]

Has the latest Plesk control panel implemented the postfix installation to start over 1000 postfix-queue processes? Also, over 200 spawn processes.

Why does it bother me:

1/ Our monitoring system normally sends us a warning if over 250 processes are running on a busy server... this is a very new server with hardly any email going to it and there is already over 1500 procs.

2/ Don't processes take up resources? Memory, CPU time when spawning etc?

3/ Clutters up the ps listing (not a big problem I know)

4/ A lightly utilised server with software on it that spawns 1000s of processes usually points to a problem with the software or configuration

So lightly utilised in fact that the following command returns no connections: netstat -tupan | grep ESTA | grep ":25"

 

[Adrian Garcia]

jacko@ the "queue file write error" will not appear in the logs, you will see it in postmaster's inbox, there will be a message per error.

And yes, this processes are related to http://forum.parallels.com/showthread.php?p=405738

 

[jacko@]

Hi Adrian,

Thanks for your message. It appears that they have no solution for that then.

I've got 4000 processes now, so, I suspect it's a good idea to change over to qmail and see what bugs that presents.

Thanks,
jacko

 

[IgorG]

Hi Adrian,

Thanks for your message. It appears that they have no solution for that then.

I've got 4000 processes now, so, I suspect it's a good idea to change over to qmail and see what bugs that presents.

Thanks,
jacko

Right, problem still under developer's investigation and fix in testing stage now. I hope that fix will be ready very soon.