• Type: Bug
• Resolution: Fixed
• Priority: Major
• Component/s: core
• Labels:

  • crash
  • jenkins
• Environment:

  Jenkins 1.488 - Aix 5.3 - IBM Java 6

1. 

   dominator-tree-leak-suspects.PNG

   115 kB

   2012-11-16 09:23

2. 

   gc-log.PNG

   70 kB

   2012-11-16 09:23

3. 

   leak-supects-details.PNG

   28 kB

   2012-11-16 09:23

4. 

   leak-supects-graph.PNG

   24 kB

   2012-11-16 09:23

5. 

   loaded-classes-count.PNG

   43 kB

   2012-11-16 12:24

6. 

   non-heap-memory.PNG

   43 kB

   2012-11-16 12:24

[JENKINS-15837] Jenkins takes a lot of time to respond after 2 days

Collapse comment: Michael Pailloncy added a comment - 2012-11-16 12:24, Edited by Michael Pailloncy - 2012-11-16 12:26

Michael Pailloncy added a comment - 2012-11-16 12:24 - edited

I've added monitoring diagrams of loaded classes and non heap memory.
Jenkins has planted this morning at 9am.
is this related?

Expand comment: Michael Pailloncy added a comment - 2012-11-16 12:24, Edited by Michael Pailloncy - 2012-11-16 12:26

Michael Pailloncy added a comment - 2012-11-16 12:24 - edited I've added monitoring diagrams of loaded classes and non heap memory. Jenkins has planted this morning at 9am. is this related?

Collapse comment: Michael Pailloncy added a comment - 2012-11-16 17:25

Michael Pailloncy added a comment - 2012-11-16 17:25

I did an experience in running a Groovy script every minute doing :

Jenkins.getInstance().reload()

Jenkins has crashed after 2:30 with similar symptoms (see loaded-classes-count.png and non-heap-memory.png)
Maybe a problem on unloading classes ?

Expand comment: Michael Pailloncy added a comment - 2012-11-16 17:25

Michael Pailloncy added a comment - 2012-11-16 17:25 I did an experience in running a Groovy script every minute doing : Jenkins.getInstance().reload() Jenkins has crashed after 2:30 with similar symptoms (see loaded-classes-count.png and non-heap-memory.png) Maybe a problem on unloading classes ?

Collapse comment: Michael Pailloncy added a comment - 2013-01-12 09:01

Michael Pailloncy added a comment - 2013-01-12 09:01

We've switch to a new architecture of Jenkins based on a RedHat OS on the master and this problem does not appear anymore.
The OOME could be caused by a memory leak due to AIX or IBM JDK.
I'll close the ticket because nobody seems to have had the same problem.

Expand comment: Michael Pailloncy added a comment - 2013-01-12 09:01

Michael Pailloncy added a comment - 2013-01-12 09:01 We've switch to a new architecture of Jenkins based on a RedHat OS on the master and this problem does not appear anymore. The OOME could be caused by a memory leak due to AIX or IBM JDK. I'll close the ticket because nobody seems to have had the same problem.

Collapse comment: Michael Pailloncy added a comment - 2013-01-12 09:02, Edited by Michael Pailloncy - 2013-01-12 09:03

Michael Pailloncy added a comment - 2013-01-12 09:02 - edited

Not reproduced on a RedHat OS with Oracle JDK.

Expand comment: Michael Pailloncy added a comment - 2013-01-12 09:02, Edited by Michael Pailloncy - 2013-01-12 09:03

Michael Pailloncy added a comment - 2013-01-12 09:02 - edited Not reproduced on a RedHat OS with Oracle JDK.