Best way to migrate to new server

[depasseg]

I replicated everything from prod to backup, re-swizzled my Prod storage pool makeup, replicated from Backup to New-Prod, and then renamed new-prod to prod. I had 20+ TB, so it took a couple days (hence the use of new-prod, I didn't want any of the existing snapshots and replication tasks to fire. Little did I know that the tasks got removed once I detached the Pro pool).

I would assume you could just do the jail as long as it was it's own dataset with it's own snapshot.

 

[Spacemarine]

Sure can! It really is that easy. Just make sure you reverse the flow while the two nodes still share a common snapshot or you will have to re-replicate all of the data.

I don't understand this, especially after reading about ZFS replication for the last few hours on this forum. If I understand correctly, once you modify the filesystem on the pull side, you have to do a completely new replication. For this reason, the filesystem on the Pull side is read-only by standard. See also here: https://forums.freenas.org/index.ph...-the-destination-read-write.10551/#post-47541

So you suggest to reverse the flow, but what happens if the filesystem on the old push changed before becoming the new pull? Then it should also fail, right? In that case, wouldn't I need to make Push read-only, take the last snapshot and then reverse the flow? And then after reversing the flow, make the new Push writeable?