Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-39081

Slow network causes metal IPI bootstrap to fail

XMLWordPrintable

    • Moderate
    • None
    • 3
    • Metal Platform 259
    • 1
    • False
    • Hide

      None

      Show
      None
    • Hide
      Increased timeout of extract-machine-os bootstrap service to allow it time to download its large container image over slow connections.
      Show
      Increased timeout of extract-machine-os bootstrap service to allow it time to download its large container image over slow connections.
    • Bug Fix
    • In Progress

      If the network to the bootstrap VM is slow, the extract-machine-os.service can time out (after 180s). If this happens, it will be restarted but services that depend on it (like ironic) will never be started even once it succeeds. systemd added support for Restart:on-failure for Type:oneshot services, but they still don't behave the same way as other types of services.

      This can be simulated in dev-scripts by doing:

      sudo tc qdisc add dev ostestbm root netem rate 33Mbit

              dhiggins@redhat.com Derek Higgins
              zabitter Zane Bitter
              Jad Haj Yahya Jad Haj Yahya
              Votes:
              0 Vote for this issue
              Watchers:
              7 Start watching this issue

                Created:
                Updated: