Closed 1wolfalone1 closed 3 weeks ago
Did you also use nodegroup plugin or just multi schedulers?
As a test, I'm currently using multi-schedulers with a node selector in that example. If this approach proves successful, I'll then activate the node group later.
There is truly a probelm in versiob v1.8.2,same issue https://github.com/volcano-sh/volcano/issues/3444 , will fix it later: )
/good-first-issue
@Monokaix: This request has been marked as suitable for new contributors.
Please ensure the request meets the requirements listed here.
If this request no longer meets these requirements, the label can be removed
by commenting with the /remove-good-first-issue
command.
I am interested in this issue, can you assign it to me?
I am interested in this issue, can you assign it to me?
Thanks for your warming! It has been fixed by https://github.com/volcano-sh/volcano/pull/3594,you can still select other good first issue to contribute: )
/close
@Monokaix: Closing this issue.
What happened: I attempted to configure multiple schedulers in Volcano and encountered unexpected behavior with different versions of the volcano-scheduler image.
When I set the volcano-scheduler image to version 1.8.2 (or the default latest in the Helm deployment), I configured a nodeSelector with volcano.sh/nodegroup-name=g2, which should match two nodes: agent2 and agent3. However, the Volcano scheduler logs indicated that it was collecting all nodes in the cluster instead of just those matching the configured node selector.
This led me to believe there might be an error in my configuration. However, downgrading the volcano-scheduler image to version 1.8.1 or below (within the Helm deployment) resolved the issue. The logs confirmed this behavior, showing that only nodes agent2 and agent3 were collected, and resource calculation for the scheduler in proportion plugins functioned correctly. What you expected to happen: The ability to create multiple schedulers with a nodeSelector should function correctly from version 1.8.2 of the volcano-scheduler image onwards.
Environment:
kubectl version
): 1.30.1