zcs 7.1.2 high CPU loading used by java process

10424bofh · Post by **10424bofh** » Tue Jan 03, 2012 4:14 am

Naa still not out it just means it WILL be fixed in the 7.1.4

mrcoffee11 · Post by **mrcoffee11** » Sun Jul 01, 2012 4:21 pm

Brad46526, did you solve this issue? Got it too on Centos 6.2 Zimbra 7.2.0. Same processes going crazy five days after a fresh install. Really weird.

EDIT

Only a reboot seems to help.

babblo · Post by **babblo** » Sun Jul 01, 2012 5:35 pm

Same here... 30 users, nobody online at the moment of problem.
Physical server.

Linux 2.6.32-220.23.1.el6.x86_64 #1 SMP Mon Jun 18 18:58:52 BST 2012 x86_64 x86_64 x86_64 GNU/Linux

Release 7.2.0_GA_2669.RHEL6_64_20120410002025 CentOS6_64 FOSS edition.

Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz

8GB RAM
The process are:
32356 ? Sl 3:52 _ /opt/zimbra/java/bin/java -XX:ErrorFile=/opt/zimbra/log -client -Xmx256m -Dzimbra.home=/opt/zimbra -Djava.library.path=/opt/zimbra/lib -Djava.ext.dirs=/opt/zimbra/java/jre/lib/

ext:/opt/zimbra/lib/jars:/opt/zimbra/lib/ext-common:/opt/zimbra/lib/ext/clamscanner -Djava.io.tmpdir=/opt/zimbra/data/tmp -Dpython.cachedir=/opt/zimbra/data/tmp/jython-cachedir org.python.util.jython /opt/zi

mbra/libexec/zmconfigd
32293 ? Sl 11:38 _ /opt/zimbra/java/bin/java -Dfile.encoding=UTF-8 -server -Djava.awt.headless=true -Dsun.net.inetaddr.ttl=60 -XX:+UseConcMarkSweepGC -XX:PermSize=128m -XX:MaxPermSize=128m -XX:So

ftRefLRUPolicyMSPerMB=1 -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+PrintGCApplicationStoppedTime -XX:-OmitStackTraceInFastThrow -Xss256k -Xms1945m -Xmx1945m -Xmn486m -Djava.io.tmpdir=/opt/zi

mbra/mailboxd/work -Djava.library.path=/opt/zimbra/lib -Djava.endorsed.dirs=/opt/zimbra/mailboxd/common/endorsed -Dzimbra.config=/opt/zimbra/conf/localconfig.xml -Djetty.home=/opt/zimbra/mailboxd -DSTART=/op

t/zimbra/mailboxd/etc/start.config -jar /opt/zimbra/mailboxd/start.jar /opt/zimbra/mailboxd/etc/jetty.properties /opt/zimbra/mailboxd/etc/jetty-setuid.xml /opt/zimbra/mailboxd/etc/jetty.xml

Any idea?

babblo · Post by **babblo** » Sun Jul 01, 2012 5:49 pm

Hi, after a long search, found this:

Seems">http://www.zimbra.com/forums/administra ... 0-utc.html
Seems to work, at least with my conf.
Hope this help!

Brad46526 · Post by **Brad46526** » Fri Jul 06, 2012 5:41 am

[quote user="mrcoffee11"]Brad46526, did you solve this issue? Got it too on Centos 6.2 Zimbra 7.2.0. Same processes going crazy five days after a fresh install. Really weird.

EDIT

Only a reboot seems to help.[/QUOTE]
Mr. Coffee, no, I never found a solution. Occasionally a java process would just take off and I'd go and kill it.

I've since moved jobs and no longer have to worry about Zimbra running my CPU hot any more though.

Babblo, my problem occurred months ago, so it wasn't due to the leap second bug.

mstroven · Post by **mstroven** » Fri Dec 28, 2012 4:51 pm

[quote user="Brad46526"]Mr. Coffee, no, I never found a solution. Occasionally a java process would just take off and I'd go and kill it.

I've since moved jobs and no longer have to worry about Zimbra running my CPU hot any more though.

Babblo, my problem occurred months ago, so it wasn't due to the leap second bug.[/QUOTE]
I'm getting the same situation approximate twice per week. My machine is an amazon EC2 instance of CentOS 6.2 x86_64 with an 800GB EBS volume mounted at /opt containing Zimbra 8.0.1 and all of it's dependents.
My instance type is an "m1.medium", and I have also tried "m1.large" with identical results. I have 35 mailboxes. 99% of the time, "top" reports java at less than 20% CPU usage. However, when things start to go bad, I see multiple java processes using 90+ % cpu, and the web interface becomes completely unresponsive. Restarting with "zmcontrol restart" does not help. In fact, rebooting the VM instance does not help. The only thing I have found that consistently helps is a re-install of Zimbra. I have ZeXtras installed, but removing it did not cure my problem. Memory utilization does not seem affected. I never see > 40% memory usage (never over 20% on the bigger VM type) in "top". I am currently trying to get some meaningful data using "sar" to correlate the cpu usage spike with some underlying issue.

I have only a suspicion that EBS write-buffer performance may be the issue, and in certain instances, log writes are not happening fast enough, so everything gets slow. Is this even a possibility? Any ideas would be greatly appreciated.
-Mike

oliver2uk · Post by **oliver2uk** » Tue Jan 01, 2013 7:52 pm

Mike,
Upgrade your zimbra to 8.0.2. The java issue have been fix in that version, we experienced the same problem as you described, 8.0.2 fixes it.
Hope it helps.
Oliver

mstroven · Post by **mstroven** » Tue Jan 01, 2013 9:32 pm

Oliver,

Thanks for the update. I did the 8.0.2. upgrade yesterday morning, and updated by ZeXtras extensions as well. So far 58 hours uptime no issues yet. Will keep a close eye, but for now, "so far, so good."

oliver2uk · Post by **oliver2uk** » Tue Jan 01, 2013 10:26 pm

We have been running 8.0.2 since it came out, I think we spotted it just shortly after it was released by pure luck.
You will not have any issues from now on

mstroven · Post by **mstroven** » Wed Jan 02, 2013 1:21 pm

Sorry. NO GOOD. At 14:15 EST today, back to the same old pegged CPU at 97% usage. No one able to access web interface, Outlook and other clients unable to attach. Only recovery is by re-installing 8.0.2...