I recently swapped a bad drive and it was completely painless. My v4 system was running a RAID1 array with a collection of mostly old junk. It consisted of:
- a 4TB Western Digital Red Pro (the only new drive in the system, other than the root drive)
- a 14 year old 1TB Seagate drive,
- a couple of old 230GB drives (one SATA, one ATA)
- a 500GB SATA drive of unknown origin (I found it in a closet last summer)
This worked well until the 1TB Seagate started throwing errors. I have thought about replacing a drive with another 4TB WD Red Pro so it seemed that the time had come to do that.
I plugged in the new drive, added it to the pool, and rebalanced. That took 5 hours, which was OK as I expected it to take a while. Somewhat to my surprise, BTRFS moved all the data to the 2 4GB drives and left the others empty. I then removed the 1TB drive from the array and rebalanced; that completed by the time I refreshed the screen to check.
The only remaining task was to shut down, physically mount the drive and put the machine back together.
Kudos to Rockstor as I was able to do this entirely through the UI. Perhaps it would have been different if the drive had actually died but credit is still due for making the proactive process seamless.