Disk Failure, Accidentally rebuilt with zpool replace, disk not part of pool?

[outofspace]

So I was moving a Truenas VM from one server to another, and physically broke a drive connector. Not to worry, the pool was raidz2 and I had a cold spare. When I booted up, I was waiting on an sfp module, so I was using novnc through proxmox to get ssh session to rebuild the pool (so, no access to the web interface at the time). I used zpool replace to add the new disk (I used /dev/da1 to identify the new drive). The resilvering went well, the zpool status command shows the pool as healthy.

But when I got the network up and running, and checked the web interface under disks that da1 was not listed as being part of the pool. (It lists it as N/A Pool).

I think maybe I borked something, unless its normal for the disk to not say its part of the pool?

How can I fix this?

Output of zpool status:
pool: MainPool
state: ONLINE
scan: resilvered 166G in 01:10:41 with 0 errors on Tue Jan 11 23:27:08 2022
config:

NAME STATE READ WRITE CKSUM
MainPool ONLINE 0 0 0
raidz2-0 ONLINE 0 0 0
gptid/06f08580-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
gptid/06fe6b30-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
gptid/072902d8-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
da1 ONLINE 0 0 0
gptid/075f67f5-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
gptid/07ab9c36-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
gptid/07b4b593-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0
gptid/07b575ff-1b1f-11ec-93e9-c4e90a32f196 ONLINE 0 0 0

 

[Arwen]

TrueNAS uses a specific partition table and uses GPTIDs for the device names.
It's probable that in your rush to fix it, you missed this step. I don't have the details, but, you can use another disk's partition table as an example.

Next, the GUI does not generally know about things done to the pool or datasets from the SHELL. (However, things done with the API CLI do get recognized by the GUI.) So, export your pool and re-import your pool. That should solve the problem with associating the new disk with the pool.

 

[outofspace]

TrueNAS uses a specific partition table and uses GPTIDs for the device names.
It's probable that in your rush to fix it, you missed this step. I don't have the details, but, you can use another disk's partition table as an example.

Next, the GUI does not generally know about things done to the pool or datasets from the SHELL. (However, things done with the API CLI do get recognized by the GUI.) So, export your pool and re-import your pool. That should solve the problem with associating the new disk with the pool.

I went ahead and export/reimport from the gui. No change to the /dev/da1 disk, still shows N/A.

Is it possible I have to do another zpool replace but with the proper GPTID and just suffer through another resilver?

 

[sretalla]

Is it possible I have to do another zpool replace but with the proper GPTID and just suffer through another resilver?

Either find one of the threads where the CLI process to do it right is explained... basically wipe the disk, create a swap partition and then a zfs partition for the data and get the gptid and replace using that...

Or offline and wipe that disk in the GUI then replace in the GUI.