Hi Luke, The short answer is this:
"Standard" retry strategy uses exponential backoff and catches a wide range of rate limiting errors. The confusingly named "Default" retry strategy is an older approach which catches a smaller set of errors, and I wouldn't generally recommend using it, but it's there because people might be reliant upon its legacy behaviour for whatever reason. The "standard" and "default" naming comes from AWS. Some more info at [1] and if you want to go into down C++ AWS SDK rabbit hole on this topic, a good starting point is [2]. Hope this helps. Duncan [1] https://boto3.amazonaws.com/v1/documentation/api/latest/guide/retries.html [2] https://sdk.amazonaws.com/cpp/api/0.14.3/class_aws_1_1_client_1_1_default_retry_strategy.html On Wed, Nov 9, 2022 at 3:37 PM Luke <[email protected]> wrote: > in the docs for pyarrow.fs.S3FileSystem I see 2 options for a > retry_strategy: AwsStandardS3RetryStrategy and > AwsDefaultS3RetryStrategy. > > What is the difference between these two? > > thanks, > Luke > >
