Uploaded image for project: 'Red Hat Fuse'
  1. Red Hat Fuse
  2. ENTESB-13275

Missing Jaeger resources after OperatorHub installation

XMLWordPrintable

    • % %
    • build6
    • Hide

      Before installing the Syndesis via OperatorHub, run these two commands which create necessary resources. These commands are described here 1.2. Registering a custom resource definition for deploying Fuse Online resources
      Via operator

      ./operator install cluster
      oc project <project-name>
      ./operator -u developer
      

      Via install script

      bash install_ocp.sh --setup
      oc project <project-name>
      bash install_ocp.sh --grant developer
      

      After that you can install Fuse Online operator via OperatorHub.

      After that, when you are creating a Syndesis instance, the Jaeger is enabled by default. Because of that, you may see in the OpenShift console that jaeger-operator pod is still in the `CrashLoopBackOff` state.
      To fix that, you have to link syndesis-pull-secret with jaeger-operator service account. To restart jaeger pod, just delete it and operator creates it again.

      oc secrets link jaeger-operator syndesis-pull-secret --for=pull
      oc delete pod jaeger-operator-<ID>
      

      Do the same operations for the syndesis-jaeger pod in case it is in the error state either.

      oc secrets link syndesis-jaeger-ui-proxy syndesis-pull-secret --for=pull
      oc delete pod syndesis-jaeger
      
      Show
      Before installing the Syndesis via OperatorHub, run these two commands which create necessary resources. These commands are described here 1.2. Registering a custom resource definition for deploying Fuse Online resources Via operator ./ operator install cluster oc project <project-name> ./ operator -u developer Via install script bash install_ocp.sh --setup oc project <project-name> bash install_ocp.sh --grant developer After that you can install Fuse Online operator via OperatorHub. After that, when you are creating a Syndesis instance, the Jaeger is enabled by default. Because of that, you may see in the OpenShift console that jaeger-operator pod is still in the `CrashLoopBackOff` state. To fix that, you have to link syndesis-pull-secret with jaeger-operator service account. To restart jaeger pod, just delete it and operator creates it again. oc secrets link jaeger- operator syndesis-pull-secret -- for =pull oc delete pod jaeger- operator -<ID> Do the same operations for the syndesis-jaeger pod in case it is in the error state either. oc secrets link syndesis-jaeger-ui-proxy syndesis-pull-secret -- for =pull oc delete pod syndesis-jaeger
    • Hide
      1. Open clean OCP 4.3 (I used CRC)
      2. Open OperatorHub tab, select Red Hat Integration - Fuse Online and click on Install
      3. Select project and click on Subscribe
      4. Go to the Installed Operators tab and select Fuse Online
      5. Click on Create Instance and you will see the default CR , due to ENTESB-13272, change values to
        jaeger:
              enabled: true
              clientOnly: false
        
      6. Link Pull secret
        oc secrets link jaeger-operator syndesis-pull-secret --for=pull
        
      7. If Jaeger operator still doesn't find image, just delete the pod. The operator will create it again
      8. You will see that Jaeger pod is in the `CrashLoopBackOff` state
      Show
      Open clean OCP 4.3 (I used CRC) Open OperatorHub tab, select Red Hat Integration - Fuse Online and click on Install Select project and click on Subscribe Go to the Installed Operators tab and select Fuse Online Click on Create Instance and you will see the default CR , due to ENTESB-13272 , change values to jaeger: enabled: true clientOnly: false Link Pull secret oc secrets link jaeger- operator syndesis-pull-secret -- for =pull If Jaeger operator still doesn't find image, just delete the pod. The operator will create it again You will see that Jaeger pod is in the `CrashLoopBackOff` state

      I am installing Syndesis via OperatorHub. I set Jaeger addon as enabled during installation (because of ENTESB-13272) and I link the jaeger-operator with the syndesis-pull-secret (as it is described in the Operator overview). However, the jaeger-operator is in CrashLoopBackOff state because Jaeger CRD (which are created by operator install cluster) and clusterrole/clusterrolebindings (which are created by operator grant --user developer) missing in the cluster.

      Also, the same situation will be for the resources which are used by the Public API addon.

      ClusterRole: syndesis-auth-delegator
      ClusterRoleBinding: syndesis-<project>-auth-delegator 
      

      (The user can get Jaeger CRD from the Jaeger operator in the OperatorHub too, but still, our cluster-wide resources will be missing (ENTESB-13234) )

              parichar@redhat.com Paul Richardson
              mkralik@redhat.com Matej Kralik
              Matej Kralik Matej Kralik
              Claudio Miranda
              Votes:
              0 Vote for this issue
              Watchers:
              9 Start watching this issue

                Created:
                Updated:
                Resolved: