Log detail information when hit insufficient resource

[7sunarni]

What is the problem you're trying to solve

Sometimes if the cluster don't have enough resource for pod to schedule, and the Pod condition will have this message

...
status:
  conditions:
  - lastProbeTime: null
    lastTransitionTime: null
    message: '0/3 nodes are unavailable: 1 plugin NodeAffinity predicates failed node(s)
      didn''t match Pod''s node affinity/selector, 2 Insufficient huawei.com/Ascend910.'
    reason: Unschedulable
    status: "False"
    type: PodScheduled
...

And scheduler will log some information like below

...
Predicates failed for task <default/POD_NAME> on node <NODE_NAME>: task default/POD_NAME on node NODE_NAME fit failed: Insufficient huawei.com/Ascend910
...

And we need to get pod resource limit and request, node allocate resource to find how much resource pod asked and how node left.

I think we can add some information to log or pod status, so we can quickly find the reason. e.g. change the log like this.

...
Predicates failed for task <default/POD_NAME> on node <NODE_NAME>: task default/POD_NAME on node NODE_NAME fit failed: Insufficient huawei.com/Ascend910, `which request 2 and node has 1`.
...

Describe the solution you'd like

Add more information to log

Additional context

No response

[lowang-bh]

You can check it in log.
Those detail information is changed dynamic and cannot be put in status.

[JesseStutler]

You can check this log:

volcano/pkg/scheduler/util/predicate_helper.go

Lines 61 to 62 in d5f3c49

	klog.V(4).Infof("Considering Task <%v/%v> on node <%v>: <%v> vs. <%v>",
	task.Namespace, task.Name, node.Name, task.Resreq, node.Idle)