Zmlogger and CPU Spikes

[andreychek]

Howdy,

I've been seeing huge CPU spikes on my system, largely since having upgraded to 4.0.0. It's generally 2-3 times a day, but not always at the same times. For 30-45 minutes, the CPU load will spike to a load as high as 10.00 - 15.00, making email unusable during that time.

Running a 'top', the processes at the top are usually "mysql", "zmlogger", and "perl" (in that order).

Now, I noticed some other threads in the forums that included suggestions of removing the log processing from cron... so last night I commented out the entry that runs "zmlogprocess" every few minutes. However, I woke up this morning to seeing it occur again with the same processes at the top.

I have a decent amount of hardware and a small amount of users. It's running on a dual Xeon 2.2Ghz with 3GB of RAM, and there's maybe 10-15 user accounts. This server often receives less than 2000 messages a day, peek days might take us up to 3000-4000, but that's a trivial amount of mail IMO.

The problem has occured regularly during hours where I'm certain none of our users have been doing email, so I don't think it's related to anything our users are currently doing.

In the myslow.log file, it's filled mostly with entries such as this:

# Time: 061020  9:38:49
# User@Host: zimbra[zimbra] @ localhost [127.0.0.1]
# Query_time: 59  Lock_time: 0  Rows_sent: 0  Rows_examined: 0
commit;

If anyone happens to have any suggestions of how I might be able to resolve this, that would be wonderful :-)

Thanks!
-Eric

[rhostager]

I am running 4.03 on Fedora Core 4. I am seeing zmlogprocess taking 99% of the cpu cycles. Here is what top shows:

top - 17:39:07 up 48 min, 1 user, load average: 1.04, 0.96, 0.73
Tasks: 121 total, 3 running, 118 sleeping, 0 stopped, 0 zombie
Cpu0 : 99.7% us, 0.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 0.3% us, 0.0% sy, 0.0% ni, 99.7% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 4025392k total, 1573140k used, 2452252k free, 180640k buffers
Swap: 2031608k total, 0k used, 2031608k free, 259540k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20380 zimbra 25 0 494m 492m 2128 R 99.8 12.5 9:03.54 zmlogprocess
1 root 16 0 1740 568 492 S 0.0 0.0 0:01.07 init
2 root RT 0 0 0 0 S 0.0 0.0 0:00.01 migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0
4 root RT 0 0 0 0 S 0.0 0.0 0:00.01 migration/1
5 root 34 19 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/1
6 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 events/0
7 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 events/1
8 root 16 -5 0 0 0 S 0.0 0.0 0:00.00 khelper
9 root 11 -5 0 0 0 S 0.0 0.0 0:00.00 kthread
12 root 20 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid
102 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 kblockd/0
103 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 kblockd/1
106 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 khubd
157 root 20 0 0 0 0 S 0.0 0.0 0:00.00 pdflush
158 root 15 0 0 0 0 S 0.0 0.0 0:00.02 pdflush

How do I fix this problem?

- Rob

[rhostager]

Thank goodness for the forums.

I read through all the threads dealing with the logger. After deleting the db directory at logger/db, fixing a few permissions issues, and executing libexec/zmloggerinit as user zimbra along with the fix here the logger is now working.

- Rob

[scottnelson]

Permissions issues for which directory/files ?
The /opt/zimbra/logger/db and child dirs?

Scotty

[rhostager]

I did chown -R zimbra:zimbra on /opt/zimbra/logger/*

- Rob