Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-9994

logLevel and operatorLogLevel fields do not work for "certmanager.operator"

    XMLWordPrintable

Details

    • Bug
    • Resolution: Done
    • Normal
    • 4.14.0
    • 4.13
    • cert-manager
    • Moderate
    • CFE Sprint 234, CFE Sprint 235
    • 2
    • No
    • False
    • Hide

      None

      Show
      None

    Description

      Description of problem:

      In OCP, for any CR "xxxxxx.operator cluster":
      Updating logLevel field, then the operand pods will renew and their log level will change accordingly.
      
      Updating operatorLogLevel, then the operator pod logs' verbosity will change accordingly.
      
      Take openshift-apiserver operator for example:
      For operand, update logLevel:
      $ oc patch --type=merge openshiftapiserver.operator cluster -p='{"spec":{"logLevel":"Trace"}}'
      openshiftapiserver.operator.openshift.io/cluster patched
      
      Operand pods are renewed:
      $ oc get po -n openshift-apiserver
      NAME                         READY   STATUS    RESTARTS   AGE
      apiserver-7b7c6dc9c4-75xrh   2/2     Running   0          2m11s
      apiserver-7b7c6dc9c4-dhj7d   2/2     Running   0          1m44s
      apiserver-7b7c6dc9c4-lzwnl   2/2     Running   0          18s
      
      Operand pod log level changes to -v=6, i.e. "Trace":
      $ oc get po -n openshift-apiserver apiserver-7b7c6dc9c4-lzwnl -o yaml
      ...
            exec openshift-apiserver start ... -v=6
      ...
      
      For operator, update operatorLogLevel:
      $ oc patch --type=merge openshiftapiserver.operator cluster -p='{"spec":{"operatorLogLevel":"TraceAll"}}'
      openshiftapiserver.operator.openshift.io/cluster patched
      
      $ oc get event -n openshift-apiserver-operator
      38s       Normal   OperatorLogLevelChange   namespace/openshift-apiserver-operator   Operator log level changed from "Normal" to "TraceAll"
      
      Operator pod's logs become verbose as per the Response Body's verbosity:
      $ oc logs <operator pod> -n openshift-apiserver-operator
      ...
      I0311 12:22:31.498722       1 request.go:1171] Response Body: {"kind":"Event",..."reason":"OperatorLogLevelChange","message":"Operator log level changed from \"Normal\" to \"TraceAll\"","source":{"component":"openshift-apiserver-operator-loggingsyncer"},...}
      I0311 12:22:31.500004       1 request.go:1169] Response Body:
      00000000  6b 38 73 00 0a 0f 0a 02  76 31 12 09 43 6f 6e 66  |k8s.....v1..Conf|
      00000010  69 67 4d 61 70 12 a0 08  0a cd 01 0a 05 61 75 64  |igMap........aud|
      ...
      
      
      However, for cert-manager operator, for CR "certmanager.operator":
      $ oc get certmanager.operator cluster -o yaml
      ...
      spec:
        logLevel: Normal
        managementState: Managed
        observedConfig: null
        operatorLogLevel: Normal
      ...
      
      Updating them:
      $ oc patch --type=merge certmanager.operator cluster -p='{"spec":{"logLevel":"Trace"}}'
      
      $ oc patch --type=merge certmanager.operator cluster -p='{"spec":{"operatorLogLevel":"TraceALL"}}'
      
      $ oc get certmanager.operator cluster -o yaml
      ...
      spec:
        logLevel: Trace
        managementState: Managed
        observedConfig: null
        operatorLogLevel: TraceALL
      ...
      
      Then check operand pods:
      $ oc get po -n cert-manager
      NAME                                       READY   STATUS    RESTARTS   AGE
      cert-manager-cainjector-75fcd87cd4-lh5qf   1/1     Running   0          48m
      cert-manager-d464b7449-p7ggf               1/1     Running   0          48m
      cert-manager-webhook-6c7678c88f-2dz5v      1/1     Running   0          48m
      
      They are not renewed.
      Log level is not changed:
      $ oc get po -n cert-manager -o yaml | grep '\--v'
            - --v=2
            - --v=2
            - --v=2
      
      Check operator pod, the pod's logs' verbosity is not changed.

      Version-Release number of selected component (if applicable):

      cert-manager operator of bundle version v1.10.2-18

      How reproducible:

      Always

      Steps to Reproduce:

      1. As above
      

      Actual results:

      logLevel and operatorLogLevel fields do not work for "certmanager.operator"

      Expected results:

      logLevel and operatorLogLevel fields should respectively work for operand and operator, like any other instances of "xxxxxx.operator"s in OCP as above openshiftapiserver.operator example.
      
      Otherwise these fields' presence in "oc get certmanager.operator cluster -o yaml" will confuse users.
      
      Despite users can change operand log level via https://github.com/openshift/cert-manager-operator/pull/95 .

      Attachments

        Activity

          People

            swghosh@redhat.com Swarup Ghosh
            xxia-1 Xingxing Xia
            Xingxing Xia Xingxing Xia
            Swarup Ghosh, Thejas N (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: