V4rynth

Alright, I've followed your steps and it seems like it worked. The pool now contains only 2 drives, the dockers are still working and it seems no data was lost. Perfect. Thank you very much for your help.

Oh, something that might be worth mentioning, I haven't upgraded the pool to a zpool. Do you have any advice on if I should do it before or not?

Thanks for the quick response. Yes, I used striped pool since I planned with much more space needed than it actually turned out. The drives I want to remove are nvme0n1p1 (cache 1) and nvme1n1p1 (cache 3). The two on the adapter card that I use. But that's more a QOL thing, since I have no problem switching the SSDs to different slots if it's better to remove caches 3 and 4 and keep 1 and 2 for example. The cache pool in its current configuration contains 4x the same 500Gb Crucial SSDs and is currently utilized with around 250GB of data striped pretty evenly on all 4 drives with between 58GB to 66 GB per drive. If it's possible to remove the drives without losing any data, I'd welcome your instructions very much. Thanks a lot for that offer.

Here is the diagnostics. pheonix-diagnostics-20250415-1831.zip

Hello there, I'm running into a sort of a big problem. I'm running my rig on unraid version 7.0.1 in a cache pool only with 4 m.2 ssds in a zfs pool. Now I wanted to add a gpu to the rig, but I'm using a mini-ITX mainboard and therefore have only 2 m.2 slots on the board. The other two ssds are sitting on a pcie adapter card. To install the gpu I have to get rid of the m.2 adapter card. But I already tested, I will not be able to start the array with even a single drive missing. I've searched many forum threads and videos, but I found no solution for my case with a cache pool only variant. Does anyone have an Idea on how I should go on from here and if I can decrease my zfs pool without having to reconfigure my shares and reinstall my dockers? I'm very happy for any tips, trick,s and ideas.