pop on serverXX.mydomain.com failed

flashweb

Well-Known Member
Mar 13, 2003
255
2
168
cPanel Access Level
Root Administrator
Hi,

On one of my cpanel server, i get following message many times a day


pop failed @ Wed Oct 24 12:24:37 2007. A restart was attempted automagically.
Server is not under load. Most of the time, it is less than 1, dual cpu opteron 248 cpu.

On /var/log/chkservd.log

I found


[Wed Oct 24 12:24:37 2007] Service check ....antirelayd [+]...cpsrvd [+]...exim [+]...ftpd [+]...httpd [+]...imap [+]...mysql [+]...named [+]...pop [TIMEOUT!!!! on send
-Notification => [email protected] via EMAIL [level => 1]
Restarting pop....
system: /scripts/restartsrv_cppop
]...Done

What log i should check next to find the problem for this ? I tried cpanel upgrade, but still getting this error.

Thanks,

Yujin
 

felatio

Registered
Oct 12, 2006
3
0
151
Same problem here ... any ideas ???? Anyone ... ???

I have 6 cpanel servers and its only happening on one of them , load in under 1 everytime , could it be something oveloading the popd daemon ?
 

feijao

Member
Sep 5, 2004
17
0
151
We have the same problem on at least 2 servers.
The server load average is low (1 or 2).

Did anyone got it fixed?
 

sehh

Well-Known Member
Feb 11, 2006
579
5
168
Europe
this happens because the pop server is busy when the service check is performed by "chkservd".

"chkservd" works by connecting to the pop3 port, if it can't connect it thinks that the pop3 daemon is dead and so it restarts it.

this may look "ok" at first glance, but its wrong and a truly bad implementation on behalf of the cPanel developers.

why? simply because there is a change (on some servers it happens rather frequently) that the pop3 server will be working normally but have reached the maximum connection limit. Thus, any extra connections will be dropped, including the service check performed by "chkservd".

this happens on all the processes as far as i know (pop3/imapd/ftp/etc)

for example, if you check your FTP logs you'll see something like:
([email protected]) [INFO] New connection from 127.0.0.1
([email protected]) [INFO] Logout.

thats the "chkservd" daemon doing a silly connection, which of course can lead the problem i described above on a busy server and which results in the entire process being restarted.

unless the cPanel developers find a better way to check if a process is alive or dead, the only solution to the problem is to raise the connection limits for those processes (pop3/imap can be done via the WHM).

just make sure you don't raise them too high else you'll be a target for DoS attacks when some attacker does a brute force attack on you.