I'm in the process of moving from a home file server with multiple RAID-1 volumes to one with a single large RAID-6 volume. Both are running Ubuntu Linux (14.04 LTS for the old one, 16.04 LTS for the new), with Linux software RAID. A few observations:
1. Ubuntu 18.04 LTS RAID seems to have issues (drives intermittently dropping out of arrays after reboots). After spending most of an evening trying to figure out why this was happening (with little progress), I backed off to 16.04 LTS, which does not seem to have this issue.
2. In retrospect, I could've probably saved myself a lot of time by pulling one drive of each of the old RAID-1 arrays (allowing the arrays to continue in degraded mode), moving the drives to the new server, and copying the data off. This would've avoided the need to copy all of the data over the network, and would've sped things up considerably.
3. Software RAID-6 performance is pretty impressive for streaming reads. Nearly 800 MB/sec on a 5-stripe array using 5xxx RPM "NAS" type disks.
4. Praise the Lord for the --re-add option. Saves your butt if you accidentally pull the wrong drive tray (assuming you notice immediately, and the array isn't being written to at the time). Without the ability to (in effect) say "oops, I really didn't mean to yank that drive", I would've needed to do let it do a full array rebuild.
5. Now that most of the data has been successfully transfered, a couple of the drives from the old server are being re-purposed as members of the new server's RAID-6 array to expand its size. For the drives being repurposed, I made the mistake of not wiping the first few MB (partition table and RAID superblock) before inserting them into the new server. RAID subsystem on the new server immediately attempted to assemble a pair of degraded RAID-1 arrays when it saw the drives. Then I compounded the mistake by attempting to wipe the partition tables and RAID superblocks while the RAID subsystem apparently still thought it was managing those drives (not sure why it was doing that, I had issued STOP commands on them). This sent the drives into some sort of stupid limbo state where I couldn't get the RAID subsystem to let go of them, and couldn't get it to add them to the RAID-6 array either because it thought they were still "busy". The kernel also seemed to be pissed off that I had wiped the partition table out from under it. After poking at this a bit I finally gave up and rebooted the box; everything returned to normal.
6. Reshaping a RAID-6 array to add more drives and expand the capacity is really slow; estimated completion time is over a day to add a pair of 3 TB drives to an existing 9 TB array. This is not really surprising since all of the data on the disks must be reshuffled. The array is still "usable" during the reshape, but performance is absolutely abysmal (hence the scare quotes). Yes there are ways to throttle the reshape to leave more performance available, but then you're potentially looking at multiple days for the operation to complete; I'll just let it run full speed, since I'm not planning on using it much anyway over the next couple of days.
Nostalgia isn't what it used to be.