github-actions[bot] commented on code in PR #65508:
URL: https://github.com/apache/doris/pull/65508#discussion_r3576774430
##########
be/src/util/s3_util.cpp:
##########
@@ -349,6 +349,7 @@ std::shared_ptr<io::ObjStorageClient>
S3ClientFactory::_create_azure_client(
}
Azure::Storage::Blobs::BlobClientOptions options;
+
options.Retry.StatusCodes.insert(Azure::Core::Http::HttpStatusCode::TooManyRequests);
Review Comment:
This adds Azure 429 to the SDK retry policy, but Azure reads already run
under `S3FileReader`'s provider-independent 429 retry loop.
`AzureObjStorageClient::get_object()` returns the final
`RequestFailedException` HTTP code to the caller, and if it is still 429,
`S3FileReader::read_at_impl()` sleeps and calls `get_object()` again up to
`max_s3_client_retry` times. Since this SDK policy also uses
`max_s3_client_retry`, the default can turn one throttled read into multiple
full SDK retry sequences before the outer loop gives up, while
write/list/delete paths do not have that second loop. Please keep a single
retry budget here, or make the reader loop provider-aware, before enabling SDK
429 retries for Azure.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
