ctdb-takeover: Use the takeover_run_fail_callback() in more cases
authorMartin Schwenke <martin@meltin.net>
Tue, 3 May 2016 02:52:36 +0000 (12:52 +1000)
committerAmitay Isaacs <amitay@samba.org>
Fri, 13 May 2016 11:47:17 +0000 (13:47 +0200)
commit1f0263c6d432039f7ac3b01029dcd91350829b93
treed593f00baaa73f296790949f08bb7e24ca2acee1
parent06ad1711cfb8945a579eae8f042d75f2876034e3
ctdb-takeover: Use the takeover_run_fail_callback() in more cases

Probably due to oversight, this is currently only used for the
"takeip" step.

This does consistent error handling and provides a layer of
indirection to the passed callback, so use it for "releaseip" and
"ipreallocated" steps too.

The callback data now needs to be initialised before the first
possible jump to "ipreallocated".

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
ctdb/server/ctdb_takeover.c