Not sure if it’s related to the number of snapshots or the fact that each core on my C2750 isn’t particularly fast but i’ve noticed rebalancing to be incredibly disruptive with BTRFS often eating 100% cpu (1 cpu core)
What happens during this time is the F/S is so unresponsive that access to shares will often time out before you get a chance to do anything.
I also notice it’s always the initial access to a folder/file thats a problem once the BTRFS process finishes with whatever it was previously processing as part of the balance then it will actually preform the request (providing the requesting application hasn’t actually timed out) and then reading/writing the actual file itself occurs at the normal speed.
E.g yesterday I tried to SCP a file to the BRTFS pool via the rockstor CLI and it sat there for a good minute or so before starting the transfer, but once it started it proceeded at the normal speed.
Last time this happened the balance was to go from raid5 to raid1 due to the raid5 bugs, this time it’s being rebalanced because I needed to add drives to the pool to increase available space.