Transfering config/pool to new system with 22.02.4

[JohnnyD]

So I have built a new server with 22.02.4 as the OS raid1 like the old system, the old system has no apps installed, it just runs virtual machines, 2 x windows and 5 x debian. Can I just backup the config from the old server, then move the pools over to the new system and import the config? Will all the VM's be there? Is there a better way?

JD

 

[artlessknave]

hardware is missing from the forum requirements.
the answer is generally "yes". move the disks, import the config, and you will have essentially the same system.
note that "raid1" should not be used with TrueNAS. do you mean raidz1, or do you mean mirrors?
usually the boot drive is mirrors, but since you have no hardware listed I can't even guess what your data pool is.

 

[JohnnyD]

Thank you for the reply, I have added both the hardwares now.......

 

[artlessknave]

looks fine to me, except for no mention of backups, so if anything does go wrong...

 

[JohnnyD]

I have backups of the VM's of sorts, bit not a whole backup of the pool, I have 2 14TB drives I can install on the new machine as an empty pool1, is there an easy way to then copy the whole old pool1 over to the new drives, that way I do have the original pool should anything go wrong?

 

[artlessknave]

replication. you should be taking snapshots anyway (not a backup but do provide a solid safety margin)
then you just replicate using the GUI.

 

[JohnnyD]

Is that for each of the VM's or the pool as a whole? I am trying to get my head around snapshots and replication, but its taking a while to sink in LoL. Do i need to do a snapshot on each VM, or just on the pool? Can it be done while the VM's are up, or do they need to be shut down?

 

[artlessknave]

you can, and should, just snapshot the whole the pool on a schedule. it costs nearly no space (only changes take space), and allows rollbackup, although powered on VM's snapshots might not be fully consistant. you will have to power them down to move it all anyway, and replications are done differential, so if you setup scheduled snapshots/replicattion, it'll get your 2 pools synced. then you can shut everything down, do one last snapshot/replication, which should be fairly fast, and move your hardware around.
it does make the whole process more complex, though. you might want to do some reading and planning.

 

[JohnnyD]

OK progress, I have 4 machines replicated, using a snapshot for each vm and then replicating in full to the new machine, however, two I am struggling with, its something to do with having used the 'clone vm' when I created the first (master )one, so these two vm's have names of xxxx_clone0 and xxxx_clone1, there are some sort of auto snapshots in existance with these names in them, but I cant delete them as it says they have dependancies (I cant see any), so if I make a new snapshot like I did with the others that worked, it fails with the the message ......

• Replication "xpool1/vm/xxx - pool1/vm" failed: Passive side: Connection reset by peer Active side: cannot receive: local origin for clone pool1/vms/xxx@auto-2022-12-26_10-20 does not exist.

Any help much appreciated, I have tried cloning the vm and snapshotting that, but it still errors

 

[artlessknave]

replications with mismatched dates have issues. if the pool has a snapshot for dec12, it tries to replicate all snapshots from dec 12, but if there IS no snapshot from dec 12 for a decendant...you get an error like that one
if you want to do manually, by dataset, you are going to have to do it ... manually ... by dataset.
you can exclude the VM dataset, and setup replications for each. it's not the way I would do it. I would replicate the pool, live, until it's updated, then shut down everything, do one last replication, and migrate to the new setup

unless there is something better on scale, or some quirk I don't know of?

 

[JohnnyD]

OK somehow I have managed to get those last 2 across, something was going on with those cloned snaphots stopping it from working, one odd thing, the two problematic ones show in the new system as sparse, but all my others are thick, I dont know why, is that an issues at all? Just my windows server one to go now which is the biggie. Thank you for your help, especially over Christmas etc.......

 

[artlessknave]

show in the new system as sparse, but all my others are thick, I dont know why, is that an issues at all

I am not sure. I think that's scale specific, a KVM option maybe? I am most familiar with Core, and functions that are nearly idenidical b/w scale and core (like the replications).