Details
-
Type:
Bug
-
Status: Open (View Workflow)
-
Priority:
Major
-
Resolution: Unresolved
-
Component/s: slave-squatter-plugin
-
Labels:None
-
Environment:Hudson 1.379, slave-squatter 1.1, heavy-job 1.0
-
Similar Issues:
Description
(Couldn't find slave-squatter nor heavy-job as components, so posted in core)
I think there's a problem in the cooperation between these components.
I started using them yesterday to replace a more complicated mix of "heavy-slaves".
I have one slave with 8 executors. I configured it, using slave-squatter, to reserve 3 of these during office hours, leaving 5 for use. I then noticed a job with weight 3 claiming to wait for next available executor even though it was the only job in the queue and there were no builds on the slave (or anywhere, actually).
I tried lowering the reservation of the slave to 2 (i.e. 6 executors free) which caused the build to start - but having to keep an eye on these situations is not why I run Hudson
The only issue I could find that smells a bit like this was JENKINS-7033 (Job in build queue is not executed).
Attachments
Issue Links
- is related to
-
JENKINS-7033 Job in build queue is not executed
-
- Reopened
-
-
JENKINS-8882 Slave-Squatter + Heavy-Job plugin causes many executor threads to die
-
- Resolved
-
-
JENKINS-15231 Job always wait for the next executor
-
- Resolved
-
Activity
Field | Original Value | New Value |
---|---|---|
Link | This issue is related to JENKINS-7033 [ JENKINS-7033 ] |
Link |
This issue is related to |
Attachment | load-predictor.patch [ 20275 ] |
Assignee | Kohsuke Kawaguchi [ kohsuke ] |
Link |
This issue is related to |
Affects Version/s | current [ 10162 ] |
Component/s | slave-squatter [ 16076 ] | |
Component/s | core [ 15593 ] |
Workflow | JNJira [ 137768 ] | JNJira + In-Review [ 174768 ] |
They smell a bit similar, but are maybe not?