Hi team - I'm running into errors with my cache pool and the docker.img, but can't figure out how to reformat / rebuild the parity. I promise I did a search and tried to do whatever I could before posting here, but I haven't been able to figure this out. I have two 1TB SSD drives in the btrfs pool.

 

Useful(?) Information:

• I've tried changing the format type to anything other than btrfs, but it won't allow me to do so in the Cache Drive drop-down. My only options are auto, btrfs, btrfs (encrypted).
• Had reiserfs errors on another array drive, and was able to rebuild that successfully.
• I've tried scrubbing and rebalancing several times. I run into an uncorrectable error and it stops the scrub. Rebalancing also stops mid-point.
• I've tried scrubbing and rebalancing from both the UI and the console (via PuTTY).
• I'm 100% cool wiping this and rebuilding... just can't figure that part out.

 

Process I've used (based on information from other posts I've read):

 

1. Copy files off cache drive to a disk on the array (disk20).
2. Stop the array.
3. Remove the cache drives from the array.
4. Start the array.
5. Stop the array.
6. Add the cache drives back - do not start array yet.
7. Click on the Cache Drive link http://[ServerIP]/Main/Device?name=cache.
8. Tried changing from BTRFS to BRTFS (encrypted).
9. Start the array.
10. See that nothing has changed.

Things I haven't tried:

• Using just one drive in the pool
• Formatting via console (because I'm not sure how - I've tried Google and the unRAID search)
• Praying to the Flying Spaghetti Monster that this will "just work" after a reboot (OK, I have tried that, unsuccessfully, several times)
• Physically replacing the drive(s) or cable(s). <insert standard excuse as to why I don't want to open my case if I don't have to: "It's hard to get to, but I'll do that if I have to!">

 

I'm all ears. I didn't include my log but can add it if that's necessary. I wasn't sure what information it would provide.

 

Edit: in my haste, I forgot to include the implied "Thank you for any help and time spent, even if it does not result in a success."

Edited August 10, 20205 yr by brimnac
To place common courtesies into my post, and recognize the amount of effort the community places into helping others.

If your ssd and controller support TRIM then the easiest way to COMPLETELY WIPE your ssd is to stop the array, unassign the SSD from the slots and from the console do:

blkdiscard /dev/sd#

with sd# being the name of your ssd (e.g. sda; NVMe would be something like nvme0n1)

 

Be careful because if you put in the wrong device, you might accidentally wipe the wrong drive and there's no way back.

Thanks! That worked perfect!