~openstack-charmers/charms/trusty/neutron-gateway/next

Viewing all changes in revision 193.

  • Committer: James Page
  • Date: 2016-07-12 12:02:39 UTC
  • Revision ID: james.page@ubuntu.com-20160712120239-lyvesld5221u05s9
Avoid restart races for nova-api-metadata

It's possible that the nova-api-metadata will startup during the
time that the nova-conductor processes on the nova-cloud-controller
units are still starting up, resulting in a messaging timeout which
causes the daemon to exit 0.

Upstart will restart a service in this scenario, however systemd is
configured in packaging to only restart 'on-failure' so will not
attempt to restart.

This points to two other bugs - one that a messaging timeout results
in a exit code of 0, and that the OpenStack services under systemd
behave differently to under upstart.

Install an override file for systemd based installs to mimic the
behaviour of upstart, and deal with a code logic problem in the
restart_trigger handling to ensure that the charm does at least
try to restart the nova-api-metadata service at the right points
in time.

Change-Id: Ia08b7840efa33fd301d0e2c55bb30ae1a102cbfa
Closes-Bug: 1547122

expand all expand all

Show diffs side-by-side

added added

removed removed

Lines of Context: