May I successively add a new hdd to an existing pool?

hama

New Member
Jun 13, 2024
2
0
1
I would like to transfer all my stored data from ext3/4 to zfs. May I first build a pool consisting if two drives as raid-z1, then copying the data from a third disk onto it. As the next step wipe the third disk, add the third disk to the existing pool, then copying the data from the fourth disk to the 3-disk z1-pool. Going on the same way withall the other disks? All disks have the same size.
I would like to achieve a zfs pool with raid-z1 quality. In total there are 8 Disks, 6 filled nearly to the brink at the moment plus one for the OS. The 8 disks are in a seperate 8-hdds-bay.
Thanx for the help
Guido
 
No.

A Raidz1 of two disks makes no sense, and you can't grow a raidz pool just by throwing more single disks at it.

If you have the budget, buy a single large 18TB drive and backup everything that fits onto that. You don't even mention the disk size.

If you build a ZFS mirror pool, you can expand it with +2 more disks. But when you get around the 6-disk mark, raidz2 starts to make more sense.
 
  • Like
Reactions: hama
  • To add the drive to an existing pool, you have a couple of options:
    • As a new vdev (Virtual Device): Use zpool add [pool name] [device]. This adds the drive as an independent unit, increasing capacity but not redundancy.
    • To an existing vdev for redundancy: If your pool is set up with mirroring or RAID-Z, use zpool add [existing vdev] [device] to add the drive to the vdev. This enhances redundancy but doesn’t increase storage capacity.
  • So, I may add empty new hdds to an existing vdev pool.
 
You really need to read up on zfs. If you have a 3-disk raidz1 pool, which is not recommended anymore if your disks are over 2TB, to expand it you would need to add another 3-disk vdev.

RAIDZ2 is the standard, since the odds of a 2nd disk failing during a resilver are not zero. So start with mirrors if you want to expand by +2 disks, or start with a 6-disk raidz2 and expand that by adding another 6-disk vdev.