Retry object store reads on temporary errors. #788
Merged
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Outdated suggestions cannot be applied.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.
I noticed that, when reading many parquet files from S3, one of the reads fails occasionally with a temporary error such as "connection closed before message completed". I think such transient failures are normal for S3, and the client is expected to retry the read in this case. However, opendal was not configured to perform a retry. This commit adds a retry layer to opendal config for S3 and GCS object stores with the default configuration (exponential backoff with up to 3 retries).
This solves the issue for me. It might be useful to make this behavior configurable in the future, but default settings should work most of the time.