-
Notifications
You must be signed in to change notification settings - Fork 511
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Upgrade ray version to 1.13 #969
Conversation
Sky On-prem #763 relies a lot on |
Yes, I will wait for the on-prem PR, but please let me know the exact dependencies used in the on-prem PR, when you get time. |
…nto upgrade-ray-1.13
Tested again with smoke_test, only failed with Remaining TODO: test on-prem cluster. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The back-compat tests are fantastic @Michaelvll! We can consider adding it as a file in tests/ later.
Should we reflect that Python 3.10 is supported now?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @Michaelvll! Just a few remaining questions, plus wdyt of the Python versioning question here: #969 (review)?
…nto upgrade-ray-1.13
…nto upgrade-ray-1.13
d38132e
to
d4385b9
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @Michaelvll! LGTM.
Some thought for the future: I spent some time diffing this PR's skylet/providers/ with Ray 1.13's counterparts, and checking if the diffs include our patches. This verification is probably not exact due to eyeballing. In the future, we should consider transitioning skylet/providers/ into patches as well for easier maintenance.
sky/skylet/LICENCE
Outdated
https://github.com/ray-project/ray/tree/master/python/ray/autoscaler/_private/gcp | ||
Git Revision: ef9d9df4e7454d428a958281e9de333795dccb44 | ||
https://github.com/ray-project/ray/tree/ray-1.13.0/python/ray/autoscaler/_private/gcp | ||
Git Revision: 4ce38d001dbbe09cd21c497fedd03d692b2be3e |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are these hashes correct?
returns 404. I checked that gcp/config.py
latest commit in the 1.13 branch is 6d83a3f2832fc24f63af1c1270afdd4850d701db.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ahhh, I put the latest git commit for the tag ray-1.13.0. Let me fix it.
Just fixed a submission features indirectly affected by 1.13 for Onprem, thanks @Michaelvll for most of the relevant Ray upgrades for Sky On-prem. |
Just logging stuff here; Further discussion with @Michaelvll , it turns out to be a much deeper bug and Zhanghao found a nice fix. The bug was that there is a status transition error from |
…nto upgrade-ray-1.13
…nto upgrade-ray-1.13
Tested:
|
Closes #957, #953
TODO:
Tested:
tests/run_smoke_tests.sh
(with the changes from Check long cluster names only for GCP; robustify smoke tests. #966 )sky launch -c test-stale 'echo start; sleep 200000'
;sky queue test-stale
shows the followingsky stop test-stale
;sky start test-stale
;sky queue test-stale
shows the following:sky admin deploy
;sky launch -c aws-onprem ''
: deploy onprem on instances with ray==1.13 installed.Backward Compatibility Test (Adapted from #1005):