volcano-sh
Scheduler supports outputting more detailed scheduling events
We use volcano to schedule Pods in production environments. Recently, we found that it is difficult to directly locate the cause when Pod scheduling fails, and R&D personnel often need to use the scheduler log to assist in locating the cause.
For example，we use an under-resourced queue to schedule Pods. When we describe pod events, we only can see podgroup is not ready. However, podgroup is like an internal object of volcano. We don't want our operations staff to be aware of it.
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 4m31s volcano pod group is not ready, 1 Pending, 1 minAvailable; Pending: 1 Unschedulable
Even if we drill down into the PodGroup, sometimes we still can't quickly figure out the reason. For example, we can see that queue resource quota is insufficient, but we do not know specific resource. We can see that queue is overused, we also do not known specific resource. This is not friendly to operations.
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning Unschedulable 6m18s (x12 over 6m29s) volcano 1/1 tasks in gang unschedulable: pod group is not ready, 1 Pending, 1 minAvailable; Pending: 1 Unschedulable
Warning Unschedulable 6m18s (x12 over 6m29s) volcano 1/1 tasks in gang unschedulable: pod group is not ready, 1 Pending, 1 minAvailable; Pending: 1 Unschedulable
Warning Unschedulable 6m17s (x12 over 6m28s) volcano 1/1 tasks in gang unschedulable: pod group is not ready, 1 Pending, 1 minAvailable; Pending: 1 Unschedulable
Normal Unschedulable 89s (x297 over 6m29s) volcano queue resource quota insufficient
Normal Unschedulable 89s (x297 over 6m29s) volcano queue resource quota insufficient
Normal Unschedulable 89s (x297 over 6m29s) volcano queue resource quota insufficient
what is the version of scheduler?
Sorry for late reply. I'm using release-1.11.