Question: does/could the datastore support multiple ipfs instances?

Hi there,

Would the datastore support concurrent access (ether rw or ro) to a given bucket? that is to say, can two distinct ipfs processes (be they on the same computer, or in containers, different VMs etc.) _share_ a bucket? In the js-ipfs implementation they apparently use an object within the bucket as a lockfile, but I've had a quick look at the code and don't really see why it would be the case here (tho I am no go programmer).

Any thoughts/guidance would be gratefully received!

Review the go-ds-s3 datastore implementation and its S3 bucket access assumptions first. Determine whether concurrent read/write or read-only access by multiple IPFS processes is supported, and clarify the required locking behavior before defining any implementation work.

Written by the indexing model from the issue text.