I need help removing log VDEV from a pool in TrueNAS Scale 22.12.3.3

[I_am_Venganz]

I added an SSD log vdev to a pool but for testing, I didn't see any improvements or change to my RW so I needed to take it off and use it for installing containers and apps. I could not see a way to take it off so first I took the drive offline and then reformated it for the app storage. the only issue there is once it took it offline it made my pool degraded and I don't see how to take the log vdev out of my pool in the web interface of the latest truenas scale. I have tried following instructions from the internet but none has worked, I have data on that pool and cannot lose them. please how can I remove the log vdev from my pool and get it back to a healthy state?

 

[sretalla]

You'll need to do that at the shell/CLI.

Start from zpool status

 

[I_am_Venganz]

It worked, I just wanted to add for anyone who sees this, if you setup your truenas and don't use root, the zpool command does not work without sudo. I had to sudo su and then ran the zpool commands. Also, since my log vdev was not mirrored, there was no pool name so what I did was type zpool remove <name of main pool> <drive ID from the zpool status>. My drive is now online and healthy Thank you @sretalla

 

[HoneyBadger]

Hello @I_am_Venganz

There shouldn't be any need to use the CLI to remove a log vdev under normal circumstances - use the Topology button under the Storage pane, and then select the missing drive and choose Remove as shown in the screenshot below:

There is a bug though that hides this button incorrectly; this should be corrected in 22.12.4 and later versions.

 

[sretalla]

There shouldn't be any need to use the CLI to remove a log vdev - use the Topology button under the Storage pane, and then select the missing drive and choose Remove as shown in the screenshot below:

Agree, but last time I was dealing with this, that button was missing for that person, so I was still working to that assumption (having no available test system on SCALE with a log configured for the moment to confirm that).

 

[HoneyBadger]

Agree, but last time I was dealing with this, that button was missing for that person, so I was still working to that assumption (having no available test system on SCALE with a log configured for the moment to confirm that).

Ahh, it's the RAIDZ-data-vdev bug again. Thanks for the reminder. Hopefully we've squashed it for good in 22.12.4; I'll see if I can add some additional test cases to make sure it doesn't inadvertently show up again.