Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-35542

[CAPI install] envtest.kubeconfig is not deleted when destroying cluster

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done-Errata
    • Icon: Undefined Undefined
    • None
    • 4.16, 4.17
    • Installer / openshift-installer
    • None
    • Low
    • No
    • OpenShift SPLAT - Sprint 255, Installer Sprint 257
    • 2
    • False
    • Hide

      None

      Show
      None
    • Hide
      * Previously, when installing a cluster on {aws-short} after deleting a cluster with the `./openshift-install destroy cluster` command, the installation would fail with an error stating that there may already be a running cluster. With this update, all leftover artifacts are removed when the cluster is destroyed, resulting in successful installations afterwards. (link:https://issues.redhat.com/browse/OCPBUGS-35542[*OCPBUGS-35542*])
      Show
      * Previously, when installing a cluster on {aws-short} after deleting a cluster with the `./openshift-install destroy cluster` command, the installation would fail with an error stating that there may already be a running cluster. With this update, all leftover artifacts are removed when the cluster is destroyed, resulting in successful installations afterwards. (link: https://issues.redhat.com/browse/OCPBUGS-35542 [* OCPBUGS-35542 *])
    • Bug Fix
    • Done

      Description of problem:

      After destroying cluster, there are still some files leftover in <install-dir>/.clusterapi_output
$ ls -ltra
total 1516
drwxr-xr-x. 1 fedora fedora     596 Jun 17 03:46 ..
drwxr-x---. 1 fedora fedora      88 Jun 17 06:09 .clusterapi_output
-rw-r--r--. 1 fedora fedora 1552382 Jun 17 06:09 .openshift_install.log
drwxr-xr-x. 1 fedora fedora      80 Jun 17 06:09 . 
$ ls -ltr .clusterapi_output/
total 40
-rw-r--r--. 1 fedora fedora  2335 Jun 17 05:58 envtest.kubeconfig
-rw-r--r--. 1 fedora fedora 20542 Jun 17 06:03 kube-apiserver.log
-rw-r--r--. 1 fedora fedora 10656 Jun 17 06:03 etcd.log

Then continue installing new cluster within same install dir, installer exited with error as below:
$ ./openshift-install create cluster --dir ipi-aws
INFO Credentials loaded from the "default" profile in file "/home/fedora/.aws/credentials" 
INFO Consuming Install Config from target directory 
FATAL failed to fetch Cluster: failed to load asset "Cluster": local infrastructure provisioning artifacts already exist. There may already be a running cluster 


After removing .clusterapi_output/envtest.kubeconfig, and creating cluster again, installation is continued.

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

      4.16 nightly build

      How reproducible:

      always

      Steps to Reproduce:

      1. Launch capi-based installation
2. Destroy cluster
3. Launch new cluster within same install dir

      Actual results:

      Fail to launch new cluster within the same install dir, because .clusterapi_output/envtest.kubeconfig is still there.

      Expected results:

      Succeed to create a new cluster within the same install dir

      Additional info:

       

              rh-ee-bbarbach Brent Barbachem
              jinyunma Jinyun Ma
              Jinyun Ma Jinyun Ma
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Created:
                Updated:
                Resolved: