Ensure keepalived is restarted when necessary.
authorSofer Athlan-Guyot <sathlang@redhat.com>
Tue, 8 Nov 2016 15:44:26 +0000 (16:44 +0100)
committerSofer Athlan-Guyot <sathlang@redhat.com>
Wed, 9 Nov 2016 09:50:24 +0000 (10:50 +0100)
commitb9dcee028795bb1132176df7ef3f435039cc20cb
treee2372e8545b4150f515e8cd6a9d6b7ba92fdb552
parentb11d048e55fc20a90bcf11eed0f4cea5e1b45f7c
Ensure keepalived is restarted when necessary.

If os-collect-config/config.json is updated before an upgrade/update,
then the os-net-config run will automatically erase the keepalived
managed ips.

This is a hackish way to ensure that keepalived is restarted during the
next phase in order to have the ip recreated.

It basically adds a comment line to the keepalived.conf file (making it
different than the puppet one) if it's there. This will force a puppet
restart of the keepalive service puting the ips back on the undercloud.

Change-Id: I56b706ff44ba31aa87a63f870940831ce02a6e77
Closes-Bug: #1640213
Related-Bug: #1638029
manifests/network/os_net_config.pp