Muslim_Wookie

OK. It's really clear that instead of the Unraid team fitting to ZFS, ZFS is being made to fit into the Unraid UI. I am hopeful that this is just a result of "perfection is the enemy of progress" and the implementation is intended to mature, with the current status just being a middle ground to having ZFS available from Unraid.

What's the historical reason for not allowing numbers in the names of pools?

I don't have new pools though, I have old pools that I will export from my old server and import into Unraid. But if Unraid is not happy with aggr0/1/2/3/4/5 I have a lot of mount paths to need to edit across a number of systems, know what I mean? So any word on what the path of development is in this aspect as per my original post? Not being able to name zpools with a trailing number itself may not be a big a deal but you can clearly see that the reasoning for that, being the naming of the "devices" in the GUI being "Aggr" and "Aggr #" for each disk that makes up the zpool doesn't make sense. Also I'd just like to point out that "Oh Unraid will rename your zpools as needed"... well that's not OK hey. I don't want to be rude about it but it's definitely and engineering response to a legitimate end user question 😜

Thank you for the advice. I understand how the system works now, still clunky but I get it. It's going to cause me a big problem because the disks I need to import from various different arrays are named something like "aggr0" (no prizes for anyone that works out why I use this naming scheme). Please see the screenshots below so you can see my process throughout, and maybe they might help other people experimenting with this. I will also add, for total clarity, that the disks no longer show up as unassigned disks. What we should be able to do is create a pool via the GUI, select raidz1/2/3/2+10 or whatever, and edit the settings via the GUI. I am trying to replace OpenIndiana, so I am done with CLI and also Napp-IT because it's a mess of a web page interface. Is there somewhere that lists the intended improvements regarding ZFS management within Unraid? Frankly I'd install Napp-IT and just have a quick look through that to see what would be useful. It's clear that the GUI as it is now is legacy and ZFS is being made to fit into it, because in the pool each device should not be the name of the pool + a number. I guess this is why you can't have a zpool name with a trailing number? It would make more sense to just emulate the look of the CLI output in the GUI and if there MUST be a "device" column to cater for the possiblity of having multiple pools where one might be ZFS and another might be btrfs then just make it "Parity-1" "Parity-2" "Disk-1" "Disk-2" and so on.

Pool in the GUI is only for caching right? The zpool shows up under ZFS Master plugin only.

ZFS created via cmd line, needs reimport each reboot. Unassigned Disks plugin also consistently reports the disks as unassigned. Not sure what else to say, it's pretty straight forward... Test setup is running in a VM with a bunch of VMDKs creating the zpool. I get it, this isn't meant to run as a VM but before I commit to dropping OpenIndiana after 15 years and switching to Unraid I must get a feel for it. Also, I cannot create shares with the zpool mounted, because the SMB section says you can't create shares unless an array is online. But it is online. Overall the way of managing zpools (not arrays made of individually formatted zfs disks) seems kludgy and incomplete. Is there a GUI that handles the zpools correctly without needing to go to CLI? After 15 years of experience with it, I don't want to be working like that at home. That's my main reason for moving from OpenIndiana - easier upkeep with a larger community.