[15/n] [torch/elastic] Introduce _RendezvousStateHolder #56538
Closed
Conversation
This file contains hidden or 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
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
cbalioglu
requested review from
H-Huang,
mingzhe09088,
mrshenli,
pritamdamania87,
rohan-varma,
wayi1 and
zhaojuanmao
as code owners
facebook-github-bot
added
cla signed
oncall: distributed
💊 CI failures summary and remediationsAs of commit e17705c (more details on the Dr. CI page):
🕵️ 2 new failures recognized by patternsThe following CI failures do not appear to be due to upstream breakages:
|
This was referenced Apr 20, 2021
cbalioglu
added a commit
that referenced
this pull request
Apr 20, 2021
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) ghstack-source-id: 126985728 Pull Request resolved: #56538
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
cbalioglu
added a commit
that referenced
this pull request
Apr 21, 2021
Pull Request resolved: #56538 This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. ghstack-source-id: 127046689 Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/)
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
cbalioglu
added a commit
that referenced
this pull request
Apr 21, 2021
Pull Request resolved: #56538 This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. ghstack-source-id: 127078863 Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/)
aivanou
reviewed
Apr 25, 2021
|
|
||
| response = self.backend.set_state(state_bits, self._token) | ||
| else: | ||
| if self.cache_duration > 0: |
There was a problem hiding this comment.
I don't think we need cache, it would just complicate the rendezvous.
How is it going to be used? - is it going to be periodically invoked in a daemon thread?
There was a problem hiding this comment.
Caching (by default 1 second) actually helps quite a bit if you have separate components in your system that all use the same _RendezvousStateHolder instance to retrieve state data (as in `DynamicRendezvousHandler). As they are decoupled from each other, if they frequently query the state, this reduces the amount of traffic.
| dead_nodes = [ | ||
| node | ||
| for node, last_keep_alive in self.state.last_keep_alives.items() | ||
| if last_keep_alive < expire_time |
There was a problem hiding this comment.
According to code, we define a node dead if the latest state contains a node that tried to update(or get?) state earlier than now()-last_keep_alive . Should we rename "last_keep_alive" to something like "last_heartbeat_time"?
There was a problem hiding this comment.
See the last revision of this PR. I renamed "keep-alive" to "hearbeat" in most places.
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
This was referenced Apr 28, 2021
This was referenced Apr 28, 2021
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
cbalioglu
changed the title
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
This PR introduces the `_RendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job. Differential Revision: [D27892600](https://our.internmc.facebook.com/intern/diff/D27892600/) [ghstack-poisoned]
aivanou
approved these changes
May 3, 2021
|
This pull request has been merged in 76bccfb. |
krshrimali
pushed a commit
to krshrimali/pytorch
that referenced
this pull request
May 19, 2021
Summary: Pull Request resolved: pytorch#56538 This PR introduces the `_RendezvousStateHolder` interface and its accompanying `_BackendRendezvousStateHolder` type that is responsible for synchronizing the local rendezvous state with the other nodes. ghstack-source-id: 127684796 Test Plan: Run the existing and new unit tests. Reviewed By: tierex Differential Revision: D27892600 fbshipit-source-id: a55d884a1f9b0d742787be4dff4271e076c08962
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Stack from ghstack:
This PR introduces the
_RendezvousStateHoldertype that is responsible for synchronizing the local rendezvous state via the backend with the other nodes in the job.Differential Revision: D27892600