-
Bug
-
Resolution: Unresolved
-
Major
-
None
Issue: When a job is running on a template from EC2 and then you go and change or delete that template it causes a file leak of files `($INSTANCE_ID)/slave.log.*` which are the traditional log files open for agents but they are not cleaned up ever afterwards and as the agent is running it opens more and more of these file handles until eventually the whole instance crashes or the build is done (the files are not released at that point but they do not continue to increase)
Steps to reproduce:
1. Create a EC2 cloud
2. Create a template within that cloud
3. Create a job which takes a long period of time and uses that template
4. Run that job
5. Delete the template
6. Observe the file handles continue to increase and you can find many file handles of the format above
What I would normally expect:
I think the file handles being opened is OK but they should not continue to re-open. Also they should be cleaned up when the job is done and the agent goes away
I have not been able to reproduce locally with the code using hpi:run as of yet so I am not sure what is causing the leak
raihaan out of curiosity, have you seen this issue reproduced in any other way after https://github.com/jenkinsci/ec2-plugin/pull/429 fixed the NPE situation?