Chris Lu

No. This would need more coding to distribute the workload.

The log said `/mnt/disk/weed` does not exist and I trust the log. :)

is it run inside a docker container? Maybe the path is not mounted?

Do you prefer to downgrade the versioning system to be in sync with the SeaweedFS versions?

try to add some debug message in these places: * https://github.com/seaweedfs/seaweedfs/blob/master/weed/command/backup.go#L150 * https://github.com/seaweedfs/seaweedfs/blob/master/weed/storage/volume_backup.go#L66

Correct. Older volume versions do not work. Just copy the volumes directly, since they are readonly in newer versions.

to reclaim the space, use old-version `weed` binary to compact the volume.

this is not supported yet.

@zehweh In `weed shell`, `volume.copy` can only copy to another volume server. Copying to a local directory is not supported yet unless someone adds a PR for it.

Recently I added `TraverseBfsMetadata` gRPC function in order to quickly traverse the directory tree. The BFS traversal now is moved to the server side. This reduces the network round trip...