Possible to remove Failed Cache DEV in pool?

[Langenohl]

Hey all,

I am running TrueNas Scale and I have a data pool in it which 5 HDD's are working fine, but the intel optane cache drive failed completely. I removed the old drive since it was not recognized by truenas anymore (showing disk unavailable even when plugged in) and replaced it with a new intel optane drive. TrueNas found the new drive and I added it to the pool with no issues.

It was impossible for me to "Replace" the drive, so I just added another cache in the pool.

Now I cannot find anywhere to remove the old dead cache phantom drive from the pool, are you able to delete it so that the pool shows as healthy again?

Cheers,
-Marc

 

[Arwen]

Their are some bugs in the TrueNAS SCALE GUI. Make sure you find and post the bug with link here. If you can't find the bug in Jira, create one, post link here. These do need to be reported & fixed.

Remove it from the SHELL command line.

 

[Langenohl]

Thank you! Would you happen to know the shell command by chance? I am still fairly new to the platform.

 

[Arwen]

Well, it's something like this;

zpool remove POOL DEVICE

You can use zpool status to get the actual device's designation, (how ZFS sees the device).


One reason I sometimes discourage potential new users from using TrueNAS, (Core or SCALE, this applies to both), is that their is a LOT of under the hood trouble shooting that can be needed over the life time of a TrueNAS server. One very important item is to learn ZFS overall, and commands in specific. ZFS was designed for Data Center Enterprise usage... so it is not necessarily the easiest to use. But, ZFS can absolutely be used for home and small office purposes.

Some people just want a NAS with a perfect GUI, and never want to get "under the hood". TrueNAS can sometimes be exactly that. But, if, (when), something goes wrong, knowing a bit about ZFS and it's command line helps.

All that said, I don't expect nor suggest that new users learn ZFS right away. But, don't dismiss having to learn ZFS. Make a help file for your use. A plain text file with commands that you have found useful for TrueNAS or ZFS, and explain what that command does for your future self's reading. Keep a copy both on your NAS, (perhaps in a backup folder), and on your desktop. Add to it when you learn something new.

Good luck.

 

[Langenohl]

Hey Arwen,

Amazing, thank you so much. Yes I am new but I am learning a lot while getting this together for my own homelab. I have no issues going under the hood to tune, maintain things as necessary and "most" of the time it does not get me in trouble.

Cheers,
-Marc

 

[pan.peter]

Hi, similar issue here. GUI is missing the option to remove/detach/offline cache drives. This should be fixed. Is there already an open Git Issue?

 

[Johnny Fartpants]

Create a bug report using the link at the top of the page.

 

[HoneyBadger]

Hi, similar issue here. GUI is missing the option to remove/detach/offline cache drives. This should be fixed. Is there already an open Git Issue?

Hi @pan.peter

There are some bugfixes in already for 22.12.4 - is your data VDEV a RAIDZ by any chance?

NAS-122781 / 22.12.4 / NAS-122781 / UI does not present "Remove" options for L2ARC or SLOG when pool contains RAIDZ data vdevs (by chrisperedun) by bugclerk · Pull Request #8401 · truenas/webui

Added logic in canRemoveDisk and canRemoveVDEV to allow removal of VdevType.Cache and .Log when hasTopLevelRaidz is TRUE - these vdev types are still removable even when a top-level RAIDZ vdev exis...

github.com

 

[pan.peter]

Yes, that is the case. Good to know that is being tracked.

 

[Isma]

Same error here, it was simply disconnecting a cache nvme, with the nas turned off, reconnecting it to the same slot and the problem arose, so I exported the pool, cleared the two caches, and reimported the pool, the two nvme's cache gave fail, but the two nvme appeared to be added as new disks, so I assigned them to an existing pool, which was yours, reassigned them as cache, and here the result, from 1 step to 3 UNAVAIL and it looks like the two nvme, the same, are assigned.

 

[trygothic]

i have same problem

 

[trygothic]

Same error here, it was simply disconnecting a cache nvme, with the nas turned off, reconnecting it to the same slot and the problem arose, so I exported the pool, cleared the two caches, and reimported the pool, the two nvme's cache gave fail, but the two nvme appeared to be added as new disks, so I assigned them to an existing pool, which was yours, reassigned them as cache, and here the result, from 1 step to 3 UNAVAIL and it looks like the two nvme, the same, are assigned.



View attachment 69075

hi, have you fixed this problem?

 

[Isma]

Yes, when updating to cobia an option is enabled and allows you to delete it