Fix bug with forked S3PathHandler #16
Open
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
The S3PathHandler suffers from the same problem described here: dask/dask#1292. This commit stores the PID that the client was created for, and refreshes the client if the PID does not match (if the PathManager is now part of a different process).
facebook-github-bot
added
the
CLA Signed
|
@DavidMChan can you elaborate on the use case that exposes this error? |
|
Sure - this happens when you're training a model (We're training HuBERT) in fairseq using iopath's S3 manager. It seems like the fork which fairseq executes can cause this SSL error (probably because of the same underlying root cause as the Dask issue above). |
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Outdated suggestions cannot be applied.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.
When using the S3PathHandler in a large number of distributed pytorch dataloaders, you can get errors of the form:
As discussed in dask/dask#1292, this is caused when the PID of the process that owns the client is not the same as the PID of the process that makes the S3 request.
This PR addresses the issue by storing the PID of the client, and refreshing the client when the PIDs no longer match.