Uploaded image for project: 'Red Hat Advanced Cluster Management'
  1. Red Hat Advanced Cluster Management
  2. ACM-18867

The CAPA pod cannot reconcile the aws credentials secret after update

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Normal Normal
    • None
    • MCE 2.8.0
    • CAPI
    • Moderate
    • None

      Description of problem:

      After toggle the capa feature in MCE then update the 
      capa-manager-bootstrap-credentials secret with aws credentials the capa pod logs raising error NoCredentialProviders and cannot reconcile the RosaControlPlane
       
      capa-controller-manager pod logs;

      I0312 20:32:09.409519       1 rosacluster_controller.go:117] "Successfully reconciled ROSACluster" controller="rosacluster" controllerGroup="infrastructure.cluster.x-k8s.io" controllerKind="ROSACluster" ROSACluster="ns-rosa-hcp/rosa-hcp-1" namespace="ns-rosa-hcp" name="rosa-hcp-1" reconcileID="e85b4f99-a087-4191-ab24-bf21051964d1" cluster="rosa-hcp-1" controlPlane="rosa-cp-1"
      E0312 20:32:10.104307       1 controller.go:324] "Reconciler error" err=<
          failed to create scope: failed to identify the AWS caller: NoCredentialProviders: no valid providers in chain. Deprecated.
              For verbose messaging see aws.Config.CredentialsChainVerboseErrors
       > controller="rosacontrolplane" controllerGroup="controlplane.cluster.x-k8s.io" controllerKind="ROSAControlPlane" ROSAControlPlane="ns-rosa-hcp/rosa-cp-1" namespace="ns-rosa-hcp" name="rosa-cp-1" reconcileID="177ba142-cbdd-4b88-9ed3-53e6ebf9916b"
      E0312 20:32:10.685638       1 controller.go:324] "Reconciler error" err=<
          failed to create scope: failed to identify the AWS caller: NoCredentialProviders: no valid providers in chain. Deprecated.
              For verbose messaging see aws.Config.CredentialsChainVerboseErrors
       > controller="rosacontrolplane" controllerGroup="controlplane.cluster.x-k8s.io" controllerKind="ROSAControlPlane" ROSAControlPlane="ns-rosa-hcp/rosa-cp-1" namespace="ns-rosa-hcp" name="rosa-cp-1" reconcileID="26590d72-6dc4-486a-bd83-93d14d6bd303"
      E0312 20:32:11.213035       1 controller.go:324] "Reconciler error" err=<
          failed to create scope: failed to identify the AWS caller: NoCredentialProviders: no valid providers in chain. Deprecated.
              For verbose messaging see aws.Config.CredentialsChainVerboseErrors
       > controller="rosacontrolplane" controllerGroup="controlplane.cluster.x-k8s.io" controllerKind="ROSAControlPlane" ROSAControlPlane="ns-rosa-hcp/rosa-cp-1" namespace="ns-rosa-hcp" name="rosa-cp-1" reconcileID="a8f286c0-c90f-4576-998b-26b30bead2b9"
      E0312 20:32:11.803468       1 controller.go:324] "Reconciler error" err=<
          failed to create scope: failed to identify the AWS caller: NoCredentialProviders: no valid providers in chain. Deprecated.
              For verbose messaging see aws.Config.CredentialsChainVerboseErrors
       > controller="rosacontrolplane" controllerGroup="controlplane.cluster.x-k8s.io" controllerKind="ROSAControlPlane" ROSAControlPlane="ns-rosa-hcp/rosa-cp-1" namespace="ns-rosa-hcp" name="rosa-cp-1" reconcileID="77addddb-acbd-4dad-84f0-8e50887acedb" 

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

      MCE v2.8.0 , CAPA v4.18

      How reproducible:

      not every time

      Steps to Reproduce:

      1. Install MCE v2.8
      2. Enable the CAPI & CAPA feature in the MultiClusterEngine engine CR
      3.  Edit the capa-manager-bootstrap-credentials secret under namespace multicluster-engine. Add data->credentials field with value of  AWS_B64ENCODED_CREDENTIALS
      4. Create RosaControlPlane with its related CRs (Cluster , ROSACluster CRs)
      5. Check the logs for the CAPA deployment pod under namespace multicluster-engine
      6. You should see logs as above and the cannot create ROSAControlPlane

      Actual results:

      Expected results:

      Additional info:

      Deleting the capa-controller-manager-xxxx pod solve the issue , the capa-controller-manager deployment will re-create the pod and things work fine.

              Unassigned Unassigned
              melserng Mohamed ElSerngawy
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated: