CPU spikes on clicking Job -> Configure

[Steven Deal]

Running 1.514, our Jenkins pid spikes the cpu to over 100% when bringing up the configure screen for a job. From clicking the link, to the page being rendered is taking roughly 1 minute 40 seconds. Setting the JENKINS_DEBUG_LEVEL to 9 does not show any more activity in the logs. I reduced the build history; but that did not help. Restarting Jenkins does seem to help this temporarily. However, looking at the avail. memory, memory used... I don't think thats the issue. Does anyone know what might affect rendering this page? Jenkins is running on a pretty beefy box.

4 Intel E5-2690 processors

MemTotal:       65890176 kB

MemFree:        19466532 kB

[Jesse Glick]

On 05/15/2013 01:13 PM, Steven Deal wrote:
> Does anyone know what might affect rendering this page?

You need to take a thread dump to diagnose. $server/threadDump, jstack, kill -QUIT, etc.

[Steven Deal]

Jstack was not that helpful for me, I'm simply not that familiar with using the tool. Looking at the thread for the 'configure' request I see (here's a stack snippet)

   java.lang.Thread.State: RUNNABLE

        at java.lang.Throwable.fillInStackTrace(Native Method)

        at java.lang.Throwable.fillInStackTrace(Throwable.java:782)

        - locked <0x00000006b6be4768> (a java.lang.reflect.InvocationTargetException)

        at java.lang.Throwable.<init>(Throwable.java:310)

        at java.lang.Exception.<init>(Exception.java:102)

        at java.lang.ReflectiveOperationException.<init>(ReflectiveOperationException.java:89)

        at java.lang.reflect.InvocationTargetException.<init>(InvocationTargetException.java:72)

        at sun.reflect.GeneratedMethodAccessor4.invoke(Unknown Source)

        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

        at java.lang.reflect.Method.invoke(Method.java:601)

        at jenkins.ClassLoaderReflectionToolkit.findClass(ClassLoaderReflectionToolkit.java:44)

        at hudson.PluginManager$UberClassLoader.findClass(PluginManager.java:951)

        at java.lang.ClassLoader.loadClass(ClassLoader.java:423)

        - locked <0x0000000413248790> (a hudson.PluginManager$UberClassLoader)

        at java.lang.ClassLoader.loadClass(ClassLoader.java:356)

        at net.java.sezpoz.IndexItem.element(IndexItem.java:134)

        at net.java.sezpoz.IndexItem.instance(IndexItem.java:178)

        at hudson.plugins.emailext.plugins.ContentBuilder.getPrivateMacros(ContentBuilder.java:79)

        at hudson.plugins.emailext.plugins.content.ScriptContent$MethodMissingHandler.methodMissing(ScriptContent.java:168)

But in the logs, I noticed the following:

WARNING: Failed to resolve class

com.thoughtworks.xstream.mapper.CannotResolveClassException: org.jenkinsci.plugins.envinject.EnvInjectPluginAction

at com.thoughtworks.xstream.mapper.DefaultMapper.realClass(DefaultMapper.java:56)

So, I removed the use of the EnvInject plugin, uninstalled it and so far... the spiking has disappeared. Course... the real test will be after many hours of uptime.

[Steven Deal]

Actually, the spiking has returned.

[Steven Deal]

This is resolved. See Jira ticket JENKINS-17910