galaxyproject/cloudman

Set master to not run jobs when autoscaling is enabled

hackdna opened this issue · 2 comments

We have the following use case: run a small master node continuously (for web UI) and spin up workers on demand (with min: 0 and max: N).

Right now (using cloudman-dev bucket) when I enable autoscaling, the master is still set to run jobs. Also, even if I set the master not to run jobs, when the last idle worker terminates the master is set to run jobs again automatically:

22:58:21 - Initiated requested termination of instance. Terminating 'i-0e0226d1db2ed616b'.
22:58:21 - Initiated requested termination of instances. Terminating '1' instances.
22:58:21 - Terminating instance i-0e0226d1db2ed616b
22:58:24 - Instance 'i-0e0226d1db2ed616b' removed from the internal instance list.
22:58:24 - The master instance is set to execute jobs. To manually change this, use the CloudMan Admin panel.

It would be great to prevent the master from running jobs by default when autoscaling is enabled.

Master won't automatically switch to being an exec host as long as autoscaling is on now. I've uploaded this code to cloudman-dev and cloudman-test buckets.

Thank you for a quick fix, I've tested this successfully using cloudman-dev. Looking forward to the next production release.