-
Notifications
You must be signed in to change notification settings - Fork 6k
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
Update KubeRay Autoscaler to use NumOfHosts for min/max workers #48212
base: master
Are you sure you want to change the base?
Conversation
Signed-off-by: Ryan O'Leary <[email protected]>
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.
LGTM
@hongchaodeng is a code owner able to review/merge this for me? |
Signed-off-by: ryanaoleary <[email protected]>
Related Issue: |
Closes #2820 |
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.
please properly rebase.
Signed-off-by: Ryan O'Leary <[email protected]>
Signed-off-by: Ryan O'Leary <[email protected]>
700abff
to
0182cf4
Compare
Oh sorry about that, I fixed the improper rebase, not sure why those other commits got added to the PR. |
Signed-off-by: Ryan O'Leary <[email protected]>
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.
What happens if maxReplicas
is less than replicas * numOfHosts
? I guess the Ray Autoscaler will terminate the additional Pods?
@@ -109,7 +109,7 @@ def _get_basic_autoscaling_config() -> dict: | |||
# Same as "small-group" with a TPU resource entry added | |||
# and modified max_workers and node_config. | |||
"tpu-group": { | |||
"max_workers": 4, | |||
"max_workers": 8, |
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.
why do we need this change?
Why are these changes needed?
This PR updates
min_workers
andmax_workers
in the autoscaleravailable_node_types
to account for the value ofnumOfHosts
, defaulting to 1 when this value is not set. This doesn't block multi-host autoscaling currently, since you can just set the value ofminReplicas
andmaxReplicas
to the desired number of multi-host workers, but this change would be helpful to avoid unexpected behavior for users.Related issue number
Checks
git commit -s
) in this PR.scripts/format.sh
to lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/
under thecorresponding
.rst
file.