Busy CPU & odd SMTP activity

[iain]

This afternoon, for the first time I can remember in 18 months of running this Zimbra server, the CPU utilisation (usr+sys) jumped up to 100%. Restarting Zimbra sees things calm down for a few minutes but it's soon back at 100%.

I upgraded to 4.5.10 a week ago, but haven't seen any problems over the last week. There were no administration changes made today, but it's possible that some users logged on today for the first time since the upgrade.

Here's the process that's using all the CPU:

zimbra    1908  1907 99 17:29 ?        04:19:19 /opt/zimbra/jdk1.6.0_02/bin/java -Xms607m -Xmx607m -client -XX:NewRatio=2 -Djava.awt.headless=true -Xss256k -Dcatalina.base=/opt/zimbra/apache-tomcat-5.5.15 -Dcatalina.home=/opt/zimbra/apache-tomcat-5.5.15 -Djava.io.tmpdir=/opt/zimbra/apache-tomcat-5.5.15/temp -Djava.library.path=/opt/zimbra/lib -Djava.endorsed.dirs=/opt/zimbra/apache-tomcat-5.5.15/common/endorsed -classpath /opt/zimbra/apache-tomcat-5.5.15/bin/bootstrap.jar:/opt/zimbra/apache-tomcat-5.5.15/bin/commons-logging-api.jar:/opt/zimbra/lib/jars/zimbra-launcher.jar com.zimbra.cs.launcher.TomcatLauncher

The only thing I can see in the logs is SMPTD connections being made and dropped at a rate of one every few seconds, which isn't normal. All the connections come from the server itself.

Jan  7 21:28:03 myserver postfix/smtpd[23263]: disconnect from myserver.myhost.co.uk[xx.xx.xx.xx]
Jan  7 21:28:03 myserver postfix/smtpd[22591]: connect from myserver.myhost.co.uk[xx.xx.xx.xx]
Jan  7 21:28:03 myserver postfix/smtpd[22591]: 4C95F431829F: client=myserver.myhost.co.uk[xx.xx.xx.xx]
Jan  7 21:28:03 myserver postfix/cleanup[22982]: 4C95F431829F: message-id=<KmqrBzhSh0SG.1Yy8vsKH>

Any ideas?

[jholder]

Hi,
We're tracking an issue similar to this:
Bug 22311 - Unexplained high tomcat CPU

Are you a Zimbra Network Edition Customer?

[iain]

Hi,

I am a network edition customer. My issue may be similar, but one difference is that the "spike" lasts for hours and so far has only been fixed by restarting Zimbra.

[jholder]

Hi,

I am a network edition customer. My issue may be similar, but one difference is that the "spike" lasts for hours and so far has only been fixed by restarting Zimbra.

Please contact support and tell them that you would like them to investigate whether your server is related to that bug. They'll need to gather some stats.