hailctl
As of 0.2.15, Hail comes bundled with the new command-line utility, hailctl. Short for “Hail control”, this tool is designed to be the home for functionality that helps manage Hail deployments, and more.
You can use this functionality by installing the latest Hail (note: you must be on Python 3):
pip install -U hail
hailctl dataproc
Right now, hailctl only has one submodule, dataproc. This is a slightly-modified lift of Liam Abbott’s cloudtools, which is now deprecated. Most cloudtools usages can be moved to hailctl dataproc easily: for instance, cluster start foo becomes hailctl dataproc start foo.
Note that this functionality is experimental and may change at any time.
Future direction: support for other cloud providers
We intend to explore similar systems to hailctl dataproc for other clouds, like hailctl emr for Amazon EMR. There are existing attempts to build this functionality, like Carlos De Niz’s tool, and it may be possible to reuse them.