fix the occasion of getNodeNameAndStorageClass error message

[cupnes]

This PR fixes the following error by ignoring not found error because old events may refer the deleted pod.

1.6686630294454665e+09  ERROR   event-reconciler        failed to name of node and storage class related to the pod     {"pod": "pie-probe-10.69.1.139-ceph-canary-block-be4f7f-27807946-tc8sw", "error": "Pod \"pie-probe-10.69.1.139-ceph-canary-block-be4f7f-27807946-tc8sw\" not found"}
github.com/topolvm/pie/controllers.(*EventReconciler).SetupWithManager.func1.1
        /work/controllers/event_controller.go:170
github.com/topolvm/pie/controllers.(*EventReconciler).SetupWithManager.func1
        /work/controllers/event_controller.go:201

Signed-off-by: Yuma Ogami yuma-ogami@cybozu.co.jp
Co-authored-by: Toshikuni Fukaya toshikuni-fukaya@cybozu.co.jp

[peng225]

because old events may refer the deleted pod.

I'm not sure if this is correct.
Once the measurement for a pod finished, countedFlag[podName] is set to true. When old events come in later, the pod should be ignored by the following check, so it seems to me that the "failed to name of ..." log does not appear.

pie/controllers/event_controller.go

Lines 165 to 167 in 799d635

	if countedFlag[podName] {
	continue
	}

[toshipp]

@peng225
It can happen in the following scenario.

1. a pod was created and was deleted
2. pie captures only create event
3. controller tries to get the pod referred by the event, but it is deleted, therefore an error happens.
4. pie captures the delete event
5. controller can perform properly

[peng225]

@toshipp Thank you for your explanation. I understand.