puppet ha job timeout too close to Jenkins timeout

[Ben Nemec]

It appears that the timeout waiting for the Heat stack in the ha job is too close to the job timeout in Jenkins. In http://logs.openstack.org/75/185775/4/check-tripleo/check-tripleo-ironic-overcloud-f20puppet-ha/f19fb70/console.html that caused us to lose the logs from the third controller node because Jenkins killed the job before the logs were finished collecting. This could make it impossible to determine the cause of the time out.

2015-05-27 04:59:32.375 | + IP=192.0.2.8
2015-05-27 04:59:32.375 | + NAME=ov-d5j4eqqnczk-2-mnphqiyctnkx-Controller-lo3bm4wtz46b
2015-05-27 04:59:32.375 | + NAME=ov-d5j4eqqnczk-2-mnphqiyctnkx-Controller
2015-05-27 04:59:32.376 | + get_state_from_host ov-d5j4eqqnczk-2-mnphqiyctnkx-Controller heat-admin@192.0.2.8
2015-05-27 05:00:49.656 | Build timed out (after 175 minutes). Marking the build as failed.
2015-05-27 05:00:49.811 | Build was aborted

The logs from 192.0.2.8 were missing in the final build output.

[Ben Nemec (bnemec)]

We're setting CI job timeouts more intelligently now, so this shouldn't happen anymore.