One of the drives in my 3x 3TB raidz1 pool has gone critical, and rather than replace an OLD 3TB seagate drive, I would rather take the remaining “good” pair of 3TB wd Reds and use them as mirrors in the new pool.
I’ll also buy a pair of larger drives to mirror.
I was intending to create a new pool out of mirrors:
pool
mirror-0
3TB
3TB
mirror-1
10TB
10TB
I’d rather not discard the 3TB rd reds, as they’re “more suited” for a home nas, and I’m sure I will get some more use out of them. That said, they’ve been spinning for several years and will go bad eventually- and there’s the concern.
What happens with this pool if one of the 3TB mirror drives goes down? Does zfs support removing the mirror, downsizing from 13tb of mirror down to 10tb?
Would it be better to make 2 different pools? I’d rather not have to manage the different SMB shares and directories from a user standpoint.
Migration involves creating a new pool with the mirror of large new drives, and replicating data to it. Then you could destroy the old pool, and add the old WD Red (if they are Red Plus CMR) as a new vdev. But one would expect these drives to fail sooner than later, due to their age; simulatneous failure from both would lose the whole pool, so the question is whether you really need the extra 3 TB space and want to take the risk, rather than reusing these drives in a less critical use.
looking at my newegg purchase history “time flies” I bought these drives in 2013. They have like 50k hours online. I guess they would best serve as noted in a secondary role. at least a different pool for less critical data, or they can be redundant offline backups.
To include very old drives in your new pool is like building a new house with the best safety standards and materials, but recycling the panels from an older home to be used for the new roof.
So I’ve resolved to follow your suggestions and make a new pool out of the new disks I purchased, relegating the old drives to some non-essential purpose/homelab/tertiary backup situation.
I now have a truenas core which is configured with the old pool and there are a handful of services associated with filesystems in the pool. (SMB and webdav specifically). I would like to preserve those services and direct them at the new pool. I have the backup offline from the system, so I can just destroy the old array and make the new one. If I create the new pool and filesystems inside with the same names, will the SMB and webdav services “find them” again? Does destroying the pool automatically purge/clear the SMB and webdav service configurations?
To be safe, I’ll do a backup of the service configs so that I can restore them fresh, but maybe it’s not necessary as long as I rename everything.
