Make run-deploy.sh a bit smarter about which compute nodes are actually
authorWard Vandewege <ward@curoverse.com>
Wed, 22 Oct 2014 13:35:49 +0000 (09:35 -0400)
committerWard Vandewege <ward@curoverse.com>
Wed, 22 Oct 2014 13:35:49 +0000 (09:35 -0400)
commit1c44466e5983383785eaf25f3b5da01936cd2cad
tree6ae2a90c99bbba3e384820b30487a1db5585670c
parenta013dc7e4f0d91499d7349eee553a771551cd13e
Make run-deploy.sh a bit smarter about which compute nodes are actually
up.

Note: there are still race conditions to be addressed here. Or maybe
better, we probably just shouldn't care about compute node puppet runs
not completing succesfully. Either the node will get shut down by node
manager, or a regular hourly puppet run will DTRT.

No issue #
jenkins/run-deploy.sh