Confusing schedule result condition

[whitewindmills]

Please provide an in-depth description of the question you have: I'm confused why the ResourceBinding's condition is BindingScheduled when all clusters are filterd by filter plugins.

status:
  conditions:
  - lastTransitionTime: "2023-05-25T09:41:11Z"
    message: Failed to apply all works, see status.aggregatedStatus for details
    reason: FullyAppliedFailed
    status: "False"
    type: FullyApplied
  - lastTransitionTime: "2023-05-25T09:50:37Z"
    message: Binding has been scheduled
    reason: BindingScheduled
    status: "True"
    type: Scheduled
  schedulerObservedGeneration: 2

karmada-scheduler log:

I0525 09:50:37.081526       1 scheduler.go:336] Start to schedule ResourceBinding(default/test-pod) as scheduling type is duplicated
I0525 09:50:37.081536       1 scheduler.go:426] "Begin scheduling resource binding with ClusterAffinity" resourceBinding="default/test-pod"
I0525 09:50:37.081572       1 generic_scheduler.go:150] Cluster "member5715" is not fit, reason: cluster(s) didn't match the placement cluster affinity constraint
I0525 09:50:37.081582       1 scheduler.go:444] ResourceBinding(default/test-pod) scheduled to clusters []
I0525 09:50:37.081694       1 scheduler.go:450] "End scheduling resource binding with ClusterAffinity" resourceBinding="default/test-pod"
I0525 09:50:37.081741       1 scheduler.go:739] Begin to patch status condition to ResourceBinding(default/test-pod)
I0525 09:50:37.081839       1 event.go:294] "Event occurred" object="default/test-pod" fieldPath="" kind="ResourceBinding" apiVersion="work.karmada.io/v1alpha2" type="Warning" reason="ScheduleBindingFailed" message="0/1 clusters are available: 1 cluster(s) didn't match the placement cluster affinity constraint."
I0525 09:50:37.081859       1 event.go:294] "Event occurred" object="default/test" fieldPath="" kind="Pod" apiVersion="v1" type="Warning" reason="ScheduleBindingFailed" message="0/1 clusters are available: 1 cluster(s) didn't match the placement cluster affinity constraint."

Could someone explain to me the reason for this code design? /cc @RainbowMango @XiShanYongYe-Chang @Poor12 @chaunceyjiang

What do you think about this question?: I think the condition should be BindingFailedScheduling when all clusters are filterd by filter plugins. The code judgment here seems to be redundant. Environment:

• Karmada version:
• Kubernetes version:
• Others:

[RainbowMango]

Good question. I'll think about it. added this to v1.7 milestone, so that we won't miss it.

[whitewindmills]

Good question. I'll think about it. added this to v1.7 milestone, so that we won't miss it.

We can discuss it at the next meeting.

[RainbowMango]

Sure, that's great, please remember to add an item to the agenda.

[RainbowMango]

I see you put the agenda to Pending topics, I just moved it to Meeting 2023-06-05, please let me know if you can't present this meeting so that we can postpone it.

[whitewindmills]

I'm free to attend the meeting.

[whitewindmills]

/assign

[whitewindmills]

After community discussion, we need to introduce new scheduled condition reasons to explain the status in detail. I'll work on it. /remove-kind question /kind feature