We have a SuSE 12, sp3 server (SLES, not openSuSE). There is only one disk in use on the system, the root disk with several file systems under mounted with BTRFS.
We are seeing issues where a “btrfs balance” job starts in the middle of the production day. This appears to lock up the server, although what appears to be happening is that the balance process is consuming 100% cpu and 100% I/O.
Since we only have the one disk, and since the reason for the balancing operation is to balance block groups across all devices, do we really need to run the balancing operation at all?