how much features missing from s3 api

[vtolstov]

I'm need to some go library to proxy s3 api requests to own internal api storage. i found this project and try to understand how it suitable for my use-case.
Can you provide some info how much needs to be done to get full s3 api compatible?
P.S. I can provide pr if i decide to use this project =)

[shabbyrobe]

Ah this is a good one, we should probably map out the supported features in the README.

At the moment, all backends support the following:

• Basic object operations (get, put, delete)
• Limited listing support. Pagination is not working yet but prefix/delimiter is.
• Limited multipart upload support. Incomplete multipart uploads are stored in memory and do not survive restarts.

I have branches I'm working on at the moment for the following:

• Bucket list pagination
• Bucket Versioning
• Backend testing suite

They're coming along nicely but they're not PR-worthy yet.

Please note though: the intended use case for gofakes3 is currently for testing. It's not meant to be used for safe, persistent access to production data at the moment. There's no reason we couldn't set that as a stretch goal at a later date, but it's a long way down the road (especially while we have so much of the API left to implement - breaking changes are inevitable) and in the meantime, there are better solutions for those use cases, some of which are listed in the gofakes3 README:

• https://github.com/minio/minio
• https://github.com/andrewgaul/s3proxy

I'd be very interested to know which way you decide to go!

[gaul]

The best way to evaluate compatibility is via Ceph s3-tests.

[vtolstov]

no problem if you don't try it in production.
i'm already check rclone and minio it does not provide interface for s3 api that i can implement and get fully featured s3 like server.
so for now - you package is only one =)

[vtolstov]

if you provide some roadmap or things for which i can provide help, this is be useful for me