[SOLVED] zmlogwatch killing system

[MTecknology]

When I try to let zimbra run, zmlogwatch will eat up 100% of my RAM. I've tried reinstalling Zimbra but it continues to eat up 100% of my 4GB RAM.

I have absolutely no idea how to fix this and it really sucks because I have no mail server now.

[uxbod]

Well you could always just disable zmlogwatch ? Could you post a top etc so we can see what is happening ? Also, do you have strace installed so we can see what the process is actually doing ?

[phoenix]

Originally Posted by MTecknology

When I try to let zimbra run, zmlogwatch will eat up 100% of my RAM. I've tried reinstalling Zimbra but it continues to eat up 100% of my 4GB RAM.

I have absolutely no idea how to fix this and it really sucks because I have no mail server now.

Upgrade or new Install? Is the version in your profile the one you're currently running? When did this start and what happened prior to this problem appearing (i.e. and operating system updates)? How long does it take before it gets to 100% and does it stay there?

[MTecknology]

How do I disable that one piece? It's hard for me to really post any output because the entire system locks up and I only have remote access to it. I can try to output top to a file in a screen session, but that's probably about the best I can do.

When I try to start zimbra, it locks up at starting mailbox (iirc).

What happened right before this is that a backup failed and filled up an entire directory in /. This was a system backup, not a zimbra backup. I noticed it because I tried to install an application on the system and apt died saying there was no space left. I ran rm -rf /mnt/backup2131/ and right when it finished, things started locking up.

Sorry, I suppose this would have been good information earlier. It was a very late night. (sleep at 0500)

I can install strace but I have no idea how to use it.

[MTecknology]

Not sure if these are abnormally large or not...

root@vindico:/opt/zimbra/db/data# du -hs *
19M ibdata1
101M ib_logfile0
101M ib_logfile1
1.4M mboxgroup{any others here}
23M mboxgroup2
30M mboxgroup3

[MTecknology]

wow... strace is amazingly easy to use.

strace -u zimbra /opt/zimbra/bin/zmcontrol start 2&> log
Gives me the attached file.

I'm also piping top -u zimbra -b to a file. The following snippet was there to start but mysqld starting up doesn't seem like such a big deal considering it was during startup and died off.

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
25626 zimbra 20 0 135m 22m 4076 S 100 0.6 0:29.08 mysqld
28064 zimbra 20 0 669m 634m 2680 S 4 16.0 0:01.70 zmlogprocess

Hopefully the attached files help. I have no mail server right now and I know I'm missing a lot of important email messages.

I'm still running 'top -u zimbra -b -d 30 -n 480 > top' but things seem to magically be working better now. Even though the log process is still eating up the entire system.

I'll post an update later. Maybe Zimbra realized I'm upset with it and saw my 'Loyal Member' status and decided that it should cut me some slack. I'm not really sure what I did to earn that status (unless it's forum membership lenght), but I enjoy seeing it change.

[MTecknology]

It's definitely zmlogprocess that's causing the issue. I ran 'killall zmlogprocess' and things magically picked up speed and life became great.

It's eating up nearly all to all of my 4GB RAM which is significantly overkill from what it should be doing.

How can I flush the log cache and anything else to give it a fresh start?

[MTecknology]

Solved the issue using this Logger - Zimbra :: Wiki

Hurray! Massive and irritating issue solved.