Uploaded image for project: 'MicroShift'
  1. MicroShift
  2. USHIFT-2239

The iso boot may get stuck in CI on errors in kickstart post install section

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Major Major
    • openshift-4.16
    • None
    • None
    • None
    • 3
    • False
    • Hide

      None

      Show
      None
    • False
    • uShift Sprint 248, uShift Sprint 249

      Description of problem:

      The iso boot may get stuck in CI on errors in kickstart post install section. 
      
      The problem is that the ci_phase_iso_boot.sh uses "parallel" to execute scenarios and "parallel" expects all the child processes to exit to consider the job finished.
      
      Since we're using `--erroronfail` option in kickstart %post section, the installation is halted by anaconda displaying the %post logs on the console (probably using `more` command or its like). One needs to manually scroll until the end of %post output so that the installation continues to the eventual shut down of the system.
      
      Note: removing `--erroronfail` %post option is not possible because errors are then ignored and the process finishes successful even in case of %post script failures.

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

      4.16 and below

      How reproducible:

      100%

      Steps to Reproduce:

      1. Add "exit 1" in the %post kickstart section
      2. Run ci_phase_iso_boot.sh
      

      Actual results:

      The ci_phase_iso_boot.sh script gets stuck when errors are encountered in a VM %post section.

      Expected results:

      The ci_phase_iso_boot.sh script should finish with errors after the `--wait` timeout of virt-install call in scenario.sh is exceeeded.

            ggiguash@redhat.com Gregory Giguashvili
            ggiguash@redhat.com Gregory Giguashvili
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated:
              Resolved: