Skip to content

MCO Cluster Operator statuses frequently changing #346

@kikisdeliveryservice

Description

@kikisdeliveryservice

Unsure if this is causing any performance issues (it has not from what I can tell but)

The Cluster Operator machine-config-operator seems to cycle through statuses when the cluster is relatively idle (ie I am not performing and updates/applying mcs/etc...). Not sure why the status is changing but it makes for a confusing user experience. See the following.

$ oc get clusteroperator -n machine-config-operator yields varying statuses within very short periods of time. For example:

machine-config-operator 3.11.0-494-g18b8dc50-dirty True False False 7s
machine-config-operator 3.11.0-494-g18b8dc50-dirty False True False 1s

In another example, I see changes in the conditions: section but the bottom entry also never changes? So the top two alternate between Available, False / Progressing, True but the oldest entry also never seems to change?

$ oc get clusteroperator machine-config-operator -o yaml -w

apiVersion: config.openshift.io/v1
kind: ClusterOperator
metadata:
  creationTimestamp: 2019-01-23T18:14:33Z
  generation: 1
  name: machine-config-operator
  resourceVersion: "919947"
  selfLink: /apis/config.openshift.io/v1/clusteroperators/machine-config-operator
  uid: bbc5a98d-1f3a-11e9-a4a4-02204cc1daaa
spec: {}
status:
  conditions:
  - lastTransitionTime: 2019-01-24T18:39:17Z
    status: "False"
    type: Available
  - lastTransitionTime: 2019-01-24T18:39:17Z
    message: Progressing towards 3.11.0-494-g18b8dc50-dirty
    status: "True"
    type: Progressing
  - lastTransitionTime: 2019-01-24T02:24:05Z
    status: "False"
    type: Failing
  extension:
    master: all 3 nodes are at latest configuration master-8b907698492f2829b35f6f826511dfa6
    worker: all 3 nodes are at latest configuration worker-58b30499568a7a2588b05ed01bba91a7
  version: 3.11.0-494-g18b8dc50-dirty
apiVersion: config.openshift.io/v1
kind: ClusterOperator
metadata:
  creationTimestamp: 2019-01-23T18:14:33Z
  generation: 1
  name: machine-config-operator
  resourceVersion: "919948"
  selfLink: /apis/config.openshift.io/v1/clusteroperators/machine-config-operator
  uid: bbc5a98d-1f3a-11e9-a4a4-02204cc1daaa
spec: {}
status:
  conditions:
  - lastTransitionTime: 2019-01-24T18:39:22Z
    status: "True"
    type: Available
  - lastTransitionTime: 2019-01-24T18:39:22Z
    status: "False"
    type: Progressing
  - lastTransitionTime: 2019-01-24T02:24:05Z
    status: "False"
    type: Failing
  extension:
    master: all 3 nodes are at latest configuration master-8b907698492f2829b35f6f826511dfa6
    worker: all 3 nodes are at latest configuration worker-58b30499568a7a2588b05ed01bba91a7
  version: 3.11.0-494-g18b8dc50-dirty

I'm also not seeing anything happening in the oc logs -f for the machine-config-operator at the times indicated in the statuses.

Does anyone have an idea why this is happening? Why do the top two alternate, when from what I can tell nothing is happening in the MCO oc logs -f?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions