search

RamenDR / ramen

Apache License 2.0
70 stars 51 forks source link

S3 configuration should potentially take a bucket rather than VRG create one by itself #252

Closed ShyamsundarR closed 2 years ago

ShyamsundarR commented 2 years ago

As it stands, we require an S3 configuration and create the buckets ourselves to store the requisite cluster data.

It is my understanding that we can instead request an S3 bucket itself, and store required cluster data within. This enables S3 sharing and bucket name collision avoidance as an S3 store is used for a variety of needs (Ramen being only one of the applications using the same).

If requesting a bucket instead of access to the entire S3 store is a viable option, it looks safer as well from an access permission to the S3 store.

This issue is to analyze the same and move to requesting a bucket, rather than requesting access to the entire object store.

ShyamsundarR commented 2 years ago

Addressed in https://github.com/RamenDR/ramen/pull/320