Skip to content

Conversation

@d4l3k
Copy link
Member

@d4l3k d4l3k commented Mar 23, 2022

Summary:
This makes dns hostname lookup failures retryable since in some environments such as Kubernetes they're not guaranteed to be resolvable until the job starts. Retrying this eliminates the race condition.

Fixes #73682

Test Plan:
Added a unit test

buck test //caffe2/test/distributed:test_store

Reviewed By: aivanou

Differential Revision: D35092284

@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented Mar 23, 2022

🔗 Helpful links

💊 CI failures summary and remediations

As of commit db8033b (more details on the Dr. CI page):


None of the CI failures appear to be your fault 💚



🚧 1 fixed upstream failure:

These were probably caused by upstream breakages that were already fixed.

Please rebase on the viable/strict branch (expand for instructions)

If your commit is older than viable/strict, run these commands:

git fetch https://github.com/pytorch/pytorch viable/strict
git rebase FETCH_HEAD

This comment was automatically generated by Dr. CI (expand for details).

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

@facebook-github-bot facebook-github-bot added fb-exported oncall: distributed Add this issue/PR to distributed oncall triage queue labels Mar 23, 2022
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D35092284

Summary:
Pull Request resolved: pytorch#74641

This makes dns hostname lookup failures retryable since in some environments such as Kubernetes they're not guaranteed to be resolvable until the job starts. Retrying this eliminates the race condition.

This also fixes `sandcastle_skip_if` when used on the class instead of the method. Previously they wouldn't inherit from TestCase so just wouldn't run under buck at all.

Fixes pytorch#73682

Test Plan:
Added a unit test

```
buck test //caffe2/test/distributed:test_store
```

Reviewed By: aivanou

Differential Revision: D35092284

fbshipit-source-id: ab067f08a2cff6d6d7a17d3f1267226cb2f9293b
@d4l3k d4l3k force-pushed the export-D35092284 branch from fd5d5f8 to db8033b Compare March 23, 2022 21:46
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D35092284

@cbalioglu cbalioglu self-requested a review March 23, 2022 22:15
Copy link
Contributor

@cbalioglu cbalioglu left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

facebook-github-bot pushed a commit that referenced this pull request Mar 24, 2022
Summary:
Pull Request resolved: #74641

This makes dns hostname lookup failures retryable since in some environments such as Kubernetes they're not guaranteed to be resolvable until the job starts. Retrying this eliminates the race condition.

This also fixes `sandcastle_skip_if` when used on the class instead of the method. Previously they wouldn't inherit from TestCase so just wouldn't run under buck at all.

Fixes #73682

Test Plan:
Added a unit test

```
buck test //caffe2/test/distributed:test_store
```

Reviewed By: aivanou

Differential Revision: D35092284

fbshipit-source-id: d40bf187e52c41f551e4fe41c536b2b0015588ee
@github-actions
Copy link
Contributor

Hey @d4l3k.
You've committed this PR, but it does not have both a 'release notes: ...' and 'topics: ...' label. Please add one of each to the PR. The 'release notes: ...' label should represent the part of PyTorch that this PR changes (fx, autograd, distributed, etc) and the 'topics: ...' label should represent the kind of PR it is (not user facing, new feature, bug fix, perf improvement, etc). The list of valid labels can be found here for the 'release notes: ...' and here for the 'topics: ...'.
For changes that are 'topic: not user facing' there is no need for a release notes label.

@d4l3k d4l3k added release notes: distributed (c10d) release notes category topic: bug fixes topic category labels Mar 24, 2022
shahofblah pushed a commit that referenced this pull request Mar 25, 2022
Summary:
Pull Request resolved: #74641

This makes dns hostname lookup failures retryable since in some environments such as Kubernetes they're not guaranteed to be resolvable until the job starts. Retrying this eliminates the race condition.

This also fixes `sandcastle_skip_if` when used on the class instead of the method. Previously they wouldn't inherit from TestCase so just wouldn't run under buck at all.

Fixes #73682

Test Plan:
Added a unit test

```
buck test //caffe2/test/distributed:test_store
```

Reviewed By: aivanou

Differential Revision: D35092284

fbshipit-source-id: d40bf187e52c41f551e4fe41c536b2b0015588ee
(cherry picked from commit f890830)
@d4l3k d4l3k deleted the export-D35092284 branch July 12, 2022 22:22
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

cla signed fb-exported oncall: distributed Add this issue/PR to distributed oncall triage queue release notes: distributed (c10d) release notes category topic: bug fixes topic category

Projects

None yet

Development

Successfully merging this pull request may close these issues.

c10d/TCPStore: hostname lookup failures are fatal causing jobs to fail on Kubernetes

3 participants