Troubleshooting#

If you encounter errors while running the main playbook, there are a few things you can do:

  • Double check your variables.
  • Inspect the part that failed by opening the playbook or role at roles/role-name/tasks/main.yaml
  • Google the specific error message.
  • Re-run the role with the verbosity '-v' option to get more debugging information (more v's give more info). For example: 
ansible-playbook playbooks/setup_bastion.yaml -vvv
  • Use tags
  • To be more selective with what parts of a playbook are run, use tags.
  • To determine what part of a playbook you would like to run, open the playbook you'd like to run and find the roles parameter. Each role has a corresponding tag.
  • There are also occasionally tags for sections of a playbook or within the role themselves.
  • This is especially helpful for troubleshooting. You can add in tags under the name parameter for individual tasks you'd like to run.
  • Here's an example of using a tag:
ansible-playbook playbooks/setup_kvm_host.yaml --tags "section_2,section_3"
  • This runs only the parts of the setup_kvm_host playbook marked with tags section_2 and section_3. To use more than one tag, they must be quoted (single or double) and comma-separated (with or without spaces between).
  • E-mail Jacob Emery at jacob.emery@ibm.com
  • If it's a problem with an OpenShift verification step:
    • Open the cockpit to monitor the VMs.
    • In a web browser, go to https://kvm-host-IP-here:9090
    • Sign-in with your credentials set in the variables file
    • Enable administrative access in the top right.
    • Open the 'Virtual Machines' tab from the left side toolbar.
    • Sometimes it just takes a while, especially if it's lacking resources. Give it some time and then re-reun the playbook/role with tags.
    • If that doesn't work, SSH into the bastion as root ("ssh root@\<bastion-ip-address-here>") and then run, "export KUBECONFIG=/root/ocpinst/auth/kubeconfig" and then "oc whoami" and make sure it ouputs "system:admin". Then run the shell command from the role you would like to check on manually: i.e. 'oc get nodes', 'oc get co', etc.
    • Open the .openshift_install.log file for information on what happened and try to debug the issue.