[Flaky tests] Fix flaky rpc profiling tests #57517
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
Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/) [ghstack-poisoned]
rohan-varma
requested review from
H-Huang,
mrshenli,
pritamdamania87,
wayi1 and
zhaojuanmao
as code owners
💊 CI failures summary and remediationsAs of commit 5a4ea57 (more details on the Dr. CI page):
1 failure not recognized by patterns:
This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions to the (internal) Dr. CI Users group. |
facebook-github-bot
added
oncall: distributed
rohan-varma
added a commit
that referenced
this pull request
May 3, 2021
Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/) ghstack-source-id: 128017866 Pull Request resolved: #57517
Closes #45145 Closes #45067 Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/) [ghstack-poisoned]
rohan-varma
added a commit
that referenced
this pull request
May 4, 2021
Pull Request resolved: #57517 Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing ghstack-source-id: 128020820 Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/)
pritamdamania87
approved these changes
May 4, 2021
Closes #45145 Closes #45067 Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/) [ghstack-poisoned]
rohan-varma
added a commit
that referenced
this pull request
May 5, 2021
Pull Request resolved: #57517 Fixes the flaky tests #45145 and #45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in #43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing ghstack-source-id: 128200041 Differential Revision: [D28166602](https://our.internmc.facebook.com/intern/diff/D28166602/)
|
This pull request has been merged in 69e64b2. |
krshrimali
pushed a commit
to krshrimali/pytorch
that referenced
this pull request
May 19, 2021
Summary: Pull Request resolved: pytorch#57517 Fixes the flaky tests pytorch#45145 and pytorch#45067. The root cause is that it is not the case that all remote events will be children of the record function remote event, as other events can sometimes be profiled under the hood such as the issue described in pytorch#43868. We fix this issue by verifying that the set of events that are children on the remote end and children on the local end are the same, without necessarily enforcing specific events to be logged. Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing ghstack-source-id: 128200041 Test Plan: CI Reviewed By: pritamdamania87 Differential Revision: D28166602 fbshipit-source-id: 8145857da4642aef31f360b20db00f4328abe2ca
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:
Closes #45145
Closes #45067
Fixes the flaky tests #45145
and #45067.
The root cause is that it is not the case that all remote events will be
children of the record function remote event, as other events can sometimes be
profiled under the hood such as the issue described in
#43868.
We fix this issue by verifying that the set of events that are children on the
remote end and children on the local end are the same, without necessarily
enforcing specific events to be logged.
Tested by running the test 1000+ times and verifying it passed. Will also test on CI box before landing
Differential Revision: D28166602