I have an application with a large and growing dataset in IPFS. To prevent using double the space, I’ve been using Filestore like this: "ipfs add --nocopy ". I’m now trying to convert my application to ipfs-cluster so I’ve tried to modified my scripts that add data to use “ipfs-cluster-ctl add --nocopy” but this only seems to accept a URL. How do I specify a path and use Filestore with ipfs-cluster-ctl?
It is not possible.
I don’t remember why, though. There might be some fundamental issues. Perhaps IPFS manages filestore-adds in a way that just does not fit cluster way of adding things, or perhaps it just not implemented.
Your best shot is to add using ipfs and pin using cluster.