I'm running 4.5.9 of the Network Edition. I've been having a lot of trouble with slow performance...lots of "cancel request" dialogs in the web interface, client timeouts, etc. The server had dual quad core processors, 4 gb ram, a 1terabyte of raid 5 drives. Last week I opened a trouble ticket and one of the Zimbra techs helped me with the issue...it didn't 100% resolve it, but it seemed much more stable. The problem was that there were multiple zmlogger instances running. Performance got a bit better after the extras were stopped.
Today I was getting timeouts again. And there seems to be 4 zmlogger instances running. 2 questions:
1) Why do they keep coming back, and how to I prevent it from happening without just checking the server processes all the time.
2) how do I know which one of the 4 to kill? Also, can I just kill the process or is there a more graceful way to do it?
[root@localhost ~]# ps afx|grep zmlogger
18930 pts/0 S+ 0:00 \_ grep zmlogger
10120 ? S 10:43 \_ /usr/bin/perl /opt/zimbra/libexec/zmlogger
7897 ? S 5:17 \_ /usr/bin/perl /opt/zimbra/libexec/zmlogger
12338 ? S 0:20 \_ /usr/bin/perl /opt/zimbra/libexec/zmlogger
4531 ? S 0:01 \_ /usr/bin/perl /opt/zimbra/libexec/zmlogger