-
Committer:
MAAS Lander
-
Author(s):
Lee Trager
-
Date:
2017-04-12 18:59:01 UTC
-
mfrom:
(5921.1.9 node_side_timeout)
-
Revision ID:
maas_lander-20170412185901-gwe80eh1zkk9qy7y
[r=mpontillo][bug=1679431][author=ltrager] Add machine, region, rack side timeout checking when running commissioning and testing scripts.
Machines, regions, and racks now detect timeouts when running scripts. A buffer of one minute is added to the given timeout to allow for cleanup. If the script run time exceeds the buffered time out a signal is sent to the region that the script has timed out. The machine, region, or rack will then proceed to the next script. A timed out script will still cause the node to go transition to a failed status.
The region still checks if the script has timed out and marks the node failed. However this now happens 5 minutes after the script time out to give the machine a chance to kill the script and signal that its moving on. Region and racks are not subject to this check.