[
https://issues.jenkins-ci.org/browse/JENKINS-11920?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=159316#comment-159316
]
Stephen Morrison commented on JENKINS-11920:
--------------------------------------------
I see the same thing. I run the exact same job on Linux, HP, and Sun machines.
The job always seems to hang on Sun after completion, although for me, whilst
the job eventually disappears from the executor and is marked in the job
history as complete, when I look at the console the build output shows the "in
progress" animation circle still going.
> All jobs (even a job with no steps) hang for twenty minutes after completion
> on Solaris
> ---------------------------------------------------------------------------------------
>
> Key: JENKINS-11920
> URL: https://issues.jenkins-ci.org/browse/JENKINS-11920
> Project: Jenkins
> Issue Type: Bug
> Components: core
> Affects Versions: current
> Environment: Host: SunOS 5.10 Generic_144488-04 sun4u sparc
> SUNW,Sun-Fire
> JVM: java version "1.6.0_12"
> Java(TM) Platform, Standard Edition for Business (build 1.6.0_12-b04)
> Java HotSpot(TM) Server VM (build 11.2-b01, mixed mode)
> Run method: java -jar jenkins.war
> JVM Args: -server -d64 -XX:+UseConcMarkSweepGC -Xmx5G
> Jenkins Args: --ajp13Port=20006 --httpPort=20006
> Reporter: Simon
> Attachments: build-output-of-hanging-job.png, build-output.png,
> build-status.png, hudson.util.Process-entries.log, noopjob-configuration.png,
> thread_dumpt.txt
>
>
> We find that on our solaris host that all jobs hang for about twenty minutes
> after they have completed. This is true even of jobs that contain no steps.
> No additional plugins have been installed.
> The threaddump page reveals that the job seems to hang on reading from the
> proc file system,
> Executor #0 for master : executing noopjob #1
> "Executor #0 for master : executing noopjob #1" Id=82 Group=main RUNNABLE
> at java.io.RandomAccessFile.read(Native Method)
> at
> hudson.util.ProcessTree$Solaris$SolarisProcess.readLine(ProcessTree.java:846)
> at
> hudson.util.ProcessTree$Solaris$SolarisProcess.getEnvironmentVariables(ProcessTree.java:826)
> - locked hudson.util.ProcessTree$Solaris$SolarisProcess@6322be0b
> at
> hudson.util.ProcessTree$OSProcess.hasMatchingEnvVars(ProcessTree.java:266)
> at hudson.util.ProcessTree$Unix.killAll(ProcessTree.java:482)
> at hudson.Launcher$LocalLauncher.kill(Launcher.java:730)
> at hudson.model.AbstractBuild$AbstractRunner.run(AbstractBuild.java:484)
> at hudson.model.Run.run(Run.java:1404)
> at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:46)
> at hudson.model.ResourceController.execute(ResourceController.java:88)
> at hudson.model.Executor.run(Executor.java:230)
> The noopjob job is a "free-style software project" job and contains no steps.
> I've attached the thread dump and several screenshots
> I have more information now after increasing the logging level of
> hudson.util.ProcessTree to FINER. I have attached the log entries that were
> written during the execution of the job with no steps. From the log entries
> it seems that one or more instances of ProcessTree$Solaris$SolarisProcess are
> reading the /proc/<pid>/as of several processes (perhaps all processes on the
> host). It blocks on a few of them: 11665 for just under a minute, 20118 for
> 4 minutes, 9522 for 1 minute, 11665 for 4 minutes, and 357 for just under a
> minute. Each of these process have nothing to do with jenkins.
> I've figured out that each of these processes are instances of the same
> application - a large java application that is running on the same host.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jenkins-ci.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
