SkyPilot: Run AI and batch jobs on any infra (Kubernetes or 12+ clouds). Get unified execution, cost savings, and high GPU availability via a simple interface.
If the ray cluster is healthy, this skips the worker node ray start command, i.e. save N-1 ssh connections, each taking 2 seconds (divided by the parallelism).
This optimization comes from #4389
Tested (run the relevant ones):
[ ] Code formatting: bash format.sh
[ ] Any manual or new tests for this PR (please specify below)
If the ray cluster is healthy, this skips the worker node
ray startcommand, i.e. save N-1 ssh connections, each taking 2 seconds (divided by the parallelism).This optimization comes from #4389
Tested (run the relevant ones):
bash format.shpytest tests/test_smoke.pypytest tests/test_smoke.py::test_fill_in_the_nameconda deactivate; bash -i tests/backward_compatibility_tests.sh