-
Bug
-
Resolution: Duplicate
-
Major
-
None
-
Jenkins 1.485
Windows Server 2008 R2
JVM 1.6.0_31 with vendor == Sun
Running Jenkins as an NT service as local Administrator
OS running in a virtual machine with 2GB RAM & a few logical procs (on under-utilized host)
I am a pretty small Jenkins install. Single win2k8 R2 VM running all Jenkins roles doing 1-2 builds/hour. 2 branches. Full build execution time (which is mostly tests running) is ~6 minutes.
I have found that when under IO pressure, Jenkins fails to load historical builds for my branches. I have observed this in two scenarios:
1) Service start after boot (when OS services are still coming up)
2) If it tries to load this data when a build is execution (which is typically I/O bound) it will "forget" all of this data.
When this happens, either restarting the Jenkins service or triggering "reload config" from the manage Jenkins UI will cause it to reload this data and all is well again.
I have seen this at rare occasions in the past but it has gotten materially more frequent over the last few upgrades (I tend to stay up to date with the latest Jenkins builds, I upgrade ~weekly). Over the last couple of weeks it has become a ~daily issue, typically triggered during build execution.
If there are logs of interest let me know and I'm happy to collect them. I'm a n00b so not sure what would be helpful. I'm starting to collect diagnostic data in parallel (reading up on what would be helpful)...