-
Notifications
You must be signed in to change notification settings - Fork 29
(feat): make S3 transfer concurrency, part size, checksum configurable #61
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
(feat): make S3 transfer concurrency, part size, checksum configurable #61
Conversation
d0b2b11 to
c172bca
Compare
Added optional parameters for download and upload settings in the README.
c14e35a to
2f102dd
Compare
|
FYI: Thanks for the contribution! Just wanted to let you know we're actually moving this functionality to https://github.com/cloudfoundry/storage-cli. That one is already adding concurrency support. Maybe it will be better to first contribute there and cherry pick back to this. |
|
a-hassanin
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the changes!
Problem
The BOSH S3 CLI currently uses AWS SDK V2 defaults for S3 uploads and downloads, which are significantly lower than other CLI tools, resulting in suboptimal performance example for https://docs.cloud.google.com/distributed-cloud/docs
https://pkg.go.dev/github.com/aws/aws-sdk-go-v2/feature/s3/manager#Downloader
These conservative defaults were designed for general-purpose use cases but create a bottleneck for BOSH operations that require higher throughput. Unlike the AWS CLI, which offers tunable parameters, there was no way for users to optimize transfer performance based on their specific infrastructure requirements and network conditions.
Solution
This PR makes S3 transfer concurrency and part size configurable, allowing operators to:
These properties are purely optional and users can choose if they really need to change the default behavior. At least for google's new air-gapped offering we needed. https://docs.cloud.google.com/distributed-cloud/docs
AWS SDK Compatibility (CRC64NVME): The latest bosh-s3cli uses the newest AWS SDK V2, which defaults to x-amz-sdk-checksum-algorithm:CRC64NVME. This has caused compatibility issues to GDCH.
Solution
In the interim, we have to disable this behavior using the configurable parameter
disable_checksums.Minor improvement --> fixing the pascalcase for
RequestChecksumCalculationEnabledandUploaderRequestChecksumCalculationEnabled