Build new pool, how to migrate data?

[-fun-]

I need advice on how to best migrate data from an existing pool to a new one.

Right now I have my main pool on a single disk in one of four main drive bays. I want to create a new RAIDZ2 pool built of 4 disks and migrate all my data to this. The existing disk must be integrated into the new pool so the existing pool must be destroyed in the process.

Outline: I would first shift the system dataset to the boot volumes, stop file sharing and shut down all VMs and jails. Then I would replicate all data (recursively from the root dataset) from the existing main pool onto a secondary pool which resides on an additional backup disk in a secondary drive bay. Then I would add 3 new disks (and do decent burn in procedures). After this I would destroy the main pool and create a new pool from the old disk and the 3 new disks. After this I would replicate from the backup disk to the new pool.

I'm a bit nervous because I do not have a test system available. :-/ I will have backups of course but I have no way of backing up existing snapshots which I want to keep.

My questions regarding this process:

Can I just create the new pool with the same name of the pool I destroyed before? And can I do the replication to / from the backup pool like this (flospool is my main pool, migtank is the pool on the backup disk):

Backup:

Code:

zfs snapshot -r flospool@migrate
zfs send -R flospool@migrate | zfs receive -F migtank

Restore:

Code:

zfs send -R migtank@migrate | zfs receive -F flospool

 

[devnullius]

You can join my efforts here :) I think it will give you a lot of insights; I'm spending the next hour to finally get it working here. I never even dared to start it so wish me luck and I'll do the same for you :)

https://forums.freenas.org/index.ph...-current-pools-to-a-10tb-transfer-disk.59852/

 

[-fun-]

Thank you for the pointer. My best wishes out to you!! Seems like the discussion there tends to the zfs send / receive replication I'm planning to do also.