[Serve] Add hash function of RunningReplicaInfo #32772
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
sihanwang41
force-pushed
the
autoscaling_fix_with_RunningReplicaInfo
branch
from
0c64878
to
972a1ac
Compare
sihanwang41
changed the title
sihanwang41
force-pushed
the
autoscaling_fix_with_RunningReplicaInfo
branch
from
972a1ac
to
fae27a1
Compare
sihanwang41
force-pushed
the
autoscaling_fix_with_RunningReplicaInfo
branch
2 times, most recently
from
d54a3c1
to
78a872c
Compare
| size = "small", | ||
| srcs = serve_tests_srcs, | ||
| tags = ["exclusive", "team:serve"], | ||
| deps = [":serve_lib"], |
There was a problem hiding this comment.
This test file wasn't running in CI before? 😮
| ] | ||
| ) | ||
| ) | ||
| object.__setattr__(self, "_hash", hash_val) |
There was a problem hiding this comment.
For my education what's the benefit of this over self._hash = hash_val? If not obvious maybe add a code comment
There was a problem hiding this comment.
oh it is a hacky way to set attribute, since we use frozen for this dataclass class. (Not allow to set attribute as you mentioned)
There was a problem hiding this comment.
Makes sense! Would be good to put this in a code comment
Signed-off-by: Sihan Wang <[email protected]>
sihanwang41
force-pushed
the
autoscaling_fix_with_RunningReplicaInfo
branch
from
78a872c
to
d9902cf
Compare
Signed-off-by: Sihan Wang <[email protected]>
architkulkarni
approved these changes
Feb 23, 2023
zcin
approved these changes
Feb 24, 2023
Co-authored-by: Cindy Zhang <[email protected]> Signed-off-by: Sihan Wang <[email protected]>
|
Linkcheck failure unrelated |
architkulkarni
added
the
tests-ok
7 tasks
7 tasks
sihanwang41
added a commit
to sihanwang41/ray
that referenced
this pull request
Mar 7, 2023
When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller. Then after compute_iterable_delta is called, it will replace the same replica in in_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break the max_concurrent_queries parameter. Related issue number Closes ray-project#32652 Signed-off-by: Sihan Wang <[email protected]>
sihanwang41
added a commit
that referenced
this pull request
Mar 7, 2023
* [Serve] Add hash function of RunningReplicaInfo (#32772) When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller. Then after compute_iterable_delta is called, it will replace the same replica in in_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break the max_concurrent_queries parameter. Related issue number Closes #32652 Signed-off-by: Sihan Wang <[email protected]> * [Serve] Fix the max_concurrent_queries issue (#33022) For the `hashable` object, __eq__ and __hash__ both need to be provided for correctness. https://docs.python.org/3.9/glossary.html#term-hashable And add tests to make sure the long poll timeout issue won't happen. --------- Signed-off-by: Sihan Wang <[email protected]>
edoakes
pushed a commit
to edoakes/ray
that referenced
this pull request
Mar 22, 2023
When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller. Then after compute_iterable_delta is called, it will replace the same replica in in_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break the max_concurrent_queries parameter. Related issue number Closes ray-project#32652 Signed-off-by: Edward Oakes <[email protected]>
peytondmurray
pushed a commit
to peytondmurray/ray
that referenced
this pull request
Mar 22, 2023
When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller. Then after compute_iterable_delta is called, it will replace the same replica in in_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break the max_concurrent_queries parameter. Related issue number Closes ray-project#32652
elliottower
pushed a commit
to elliottower/ray
that referenced
this pull request
Apr 22, 2023
When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller. Then after compute_iterable_delta is called, it will replace the same replica in in_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break the max_concurrent_queries parameter. Related issue number Closes ray-project#32652 Signed-off-by: elliottower <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Why are these changes needed?
When the longpoll client timeout happens, the all internal objects will be cleaned up. The longpoll client will try to poll again. When this happens, longpoll client will receive another object which having same information but different object id (ActorHandle) from the controller.
Then after
compute_iterable_deltais called, it will replace the same replica inin_flight_queries, and cause router clean up ongoing request_ref, and keep assigning new request to the replica, which will break themax_concurrent_queriesparameter.Related issue number
Closes #32652
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.