Julia Kreger 904f44522b Be more sane about cleaning ... Seems that by default we now run cleaning on all jobs. However, this is extremely wasteful and time consuming. Also, if a cloud is IO resource constrained, it can result in jobs failing erroneously due to unpredictable IO performance, particularly with the multinode jobs. As such, slightly extend the wait timeout before we fail for nodes not being ready, and flag multinode jobs not to run cleaning. Change-Id: I12011743fa197caef08a3d6d1297d662dfa7640a		2018-01-24 12:57:05 -08:00
..
run.yaml	Be more sane about cleaning	2018-01-24 12:57:05 -08:00