Open sylvinus opened 8 years ago
I would say for now that this is out of the scope of the project, since it adds functionality that is beyond simple infrastructure management. I think you can also script this pretty easily with a few calls to flintrock copy-file
and flintrock run-command
.
However, I would like to leave this request open since this is a nice workflow improvement over doing all that stuff, and it's a common, Spark-specific thing users will want to do after launching their cluster.
(Launching spark-submit
in a screen
on the master would be great too so that you don't lose the job if disconnected)
I don't know if this is considered to be in the scope of flintrock, so this is obviously just a suggestion.
It could be handy after launching a cluster to be able to do:
This would login into the cluster and run spark-submit with the
--master
flag already configured tospark://....:7077
. Prior to that it could also maybe copy-filejobs/myjob.py
to the cluster?What do you think?