Upgrade / Replace Cache drives in zpool

[zetabax]

Afternoon community,

I'm trying to upgrade / replace my cache ssds and have ran into a bit of an issue.

Originally i had a pair of 256 GB SSDs acting as the cache pool and wanted to replace / upgrade them to a pair of newer 1 Tib SSDs.

So I pulled one of the old SSDs out and popped in the new one and expected to have the option to replace the 'failed' drive.

Unfortunately that wasn't the case. The array is on-line and unhealthy with an unrecoverable error and the 1 Tib SSD was automatically added to the pool.

I've tried the commands zpool status recommended with no results.

Guidance / suggestions is greatly appreciated please.

 

[sretalla]

So I pulled one of the old SSDs out

Did you remove it from the pool first? (either in GUI or with CLI ... zpool remove tank1 <disk or gptid>)

the 1 Tib SSD was automatically added to the pool

I'm not aware of any process that would produce that result... perhaps the way that debian does disk identification somehow made the new disk take the old identifier, but even then, your pool is showing gptids for all disks.

you might be best placed to export the pool and re-import it with the -m switch to drop the cache disks, then add back the ones you want.

 

[Etorix]

L2ARC are not essential devices so there is no need to replace them. You just remove the old device and add a new cache device to the pool.
Using L2ARC takes up RAM. 2 TB L2ARC is going to eat up an awful lot of RAM… Do you really have a working dataset that takes 2 TB and is your system already loaded up to the gills with RAM? Otherwise you may end up slowing your NAS.