charm-neutron-gateway/files/override.conf
James Page 68ea83e60d 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
2016-07-12 13:02:39 +01:00

4 lines
158 B
Plaintext

# Override default Restart policy for nova-api-metadata to always
# dealing with messaging timeout on startup with conductor services
[Service]
Restart=always