Duplicate TrueNAS Server DataSet Configuration Question

Thousandbuckle

Contributor
Joined
Jul 9, 2014
Messages
136
Hello all, I am running 2 TrueNAS 12.0-U7 servers
  • Primary server (VolumeXL0 as a 8 x 4TB disk RaidZ2 with one data set
  • Backup server (MegaPool) as a 6 x 16TB disk RaidZ2
I would like to setup a SnapShot Task then Replicate Task from (VolumeXL) to send the data to the server (MegaPool). I would like the data, jails, services, etc. to all be the same on the Backup as it is on the Target.

I am okay with the Snapshot and Replication tasks fom the primary server, my main question is how should I configure the DataSet on the Backup server (MegaPool)? Should I create a DataSet with the same name as the DataSet on the Primary server(VolumeXL) or does it matter?

Basically if my primary server VolumeXL were to take a dump from a HW or HD/Raid perspective, I could just switch over to the backup server (MegaPool) and I wouldnt miss a beat.
 
Joined
Oct 22, 2019
Messages
3,641
Dataset and snapshot wise, it's possible to get a "perfect replica" on the backup server, so to speak.

As for seamlessly not missing a beat with this setup, not so sure, since you're dealing with two distinct servers and configs, and you cannot overwrite the top-level root dataset of the target server with a Replication Task. (Unless this has changed?)

You could try with the pseudo-root method, but that will not account for the iocage and .system datasets (since TrueNAS does not allow specifying a custom hierarchy for those: they always sit directly under the top-level root dataset.)

One possibility is to create two different Replication Tasks: one for iocage and the other for your data, to nest them both under the backup server's top-level root dataset. However, due to network configurations, I don't believe it's a matter of "just start up the jails and you're good."

Hopefully someone else has tried this setup before and tested it out to confirm it works seamlessly.
 

Thousandbuckle

Contributor
Joined
Jul 9, 2014
Messages
136
I thought that if I did a SnapShot of the DataSet top level that it would capture all of the necessary information needed such as the jails etc. So in my case in the image below if I SnapShot Volume XL DataSet then replicate that to the backup server that if I had the same config from Primary copied and imported to Backup (minus the IP address) then I would have an exact copy of the one server on the other.
Screen Shot 2022-02-09 at 11.04.23 AM.png
 
Joined
Oct 22, 2019
Messages
3,641
So in my case in the image below if I SnapShot Volume XL DataSet then replicate that to the backup server that if I had the same config from Primary copied and imported to Backup (minus the IP address) then I would have an exact copy of the one server on the other.
Try it. Though I think the Replication Task might complain about how it "cannot overwrite the top-level root dataset" on the backup server.

Otherwise, it might "nest" VolumeXL under VolumeXL on the target, so that the pool layout on the backup server might look like so:
  • VolumeXL
    • VolumeXL
      • Data
      • home
      • iocage
      • multimedia
      • public

There are still configuration differences, which are housed in the "config file", that can be exported or found under the folder /data/, yet will not be touched / copied by any Replication Task.
 

Thousandbuckle

Contributor
Joined
Jul 9, 2014
Messages
136
So on my Backup Server I created a new Pool with the same name (VolumeXL) as on my Primary server. Then on the Replication task only choose the individual folders to send to the new Pool (VolumeXL) I would imagine it would be nested properly.

I created a new Pool on the Backup Server with the same Pool name as on the Primary Server (VolumeXL). I have started the replication task to for just the individual data sets from the Primary and so far the structure looks correct. It will take about 5 days to replicate the snapshot then I can shutdown the primary server and bring up the backup and see how things work.

Thank you for your inputs.
 
Joined
Oct 22, 2019
Messages
3,641
Then on the Replication task only choose the individual folders to send to the new Pool (VolumeXL) I would imagine it would be nested properly.

Other than those being "datasets" (not "folders"), sounds good! Share how it goes! But keep in mind there are differences between the servers: network, IP addresses, and other configs.
 
Top