Greetings, another newbie with questions/issues regarding upgrading a disk in a pool.  I ready many posts regarding this topic before giving it a go and still am unable to get rid of the disk I'm trying to replace.

A little history.  This is my first unraid build utilizing ten year old hardware that had recently been replace from being my primary system.  It originally had a 250 GB SATA SSD and a few mechanical disks from 3TB to 12TB in size and I originally set up with this configuration.

Almost immediately I decided to add a 2TB nvme drive to the pool.  As there was such a disparity between the two drives in the pool, I did not do any RAID configuration.  

I have now added a new 2TB SATA SSD with the intent of replacing the 250GB disk.  But btfrs doesn't seem to want to ever forget about the 250GB ten year old disk.

I had shut down docker and VM managers, moved data to the array, added the new disk, and removed the old, but the system kept complaining about the missing disk.  I even deleted the pool entirely and tried to set up a new pool of the two disks, and it would continue to complain about the missing old disk.

Eventually I decided to just get everything back up and running so recreated the pool with all three disks and restored appdata from backup and the system is up and running, but I really dislike that the entire system seems to be dependent on a 10 year old SSD.  

So any advice on how to get rid of this disk?  I really don't want to go moving to the array again if I don't have to, I don't keep anything on the pool that I don't have backed up and I can restore easily if I can just get the system to forget about this disk.

thunderstrike-diagnostics-20250505-0921.zip

If you are willing to upgrade to 7.1.0-rc.4 you can do it from the GUI, after converting the pool to raid1, with 7.0.x you will need to use the CLI, but I can post the commands if you prefer to remain on 7.0.1

I'd appreciate the commands to give that a shot first, but if necessary I will certainly try rc.4.   Thanks in advance.

First click the cache pool and balance to raid1, when that is done, and with the array still running, type:

 

btrfs device remove 1 /mnt/cache

 

When that completes, post new diags.

Running now.  Looks to be balanced pretty well already but I guess required to convert to RAID1.  

13 minutes ago, Just Bill said:

but I guess required to convert to RAID1.  

Correct.

It completed, so I stopped the array, removed the disk and started back up again.  I am now successfully able to start the array with just the two disks and everything is functioning, but btrfs is still pissed off about that disk not being there... LOL.  There are some btfs operations ongoing now that prevent me from stopping the array again, seems like more mirroring or somethign going on.  So did I jump the gun and was there more to do before I got right back into removing the disk, or is this going to work itself out when the balancing is completed again? 

 

 

thunderstrike-diagnostics-20250505-1222.zip

31 minutes ago, Just Bill said:

It completed, so I stopped the array, removed the disk and started back up again. 

I didn't ask for that, but in this case it shouldn't hurt, on main click on the first device for that pool and then "remove pool"
back on main, create a new pool with the same name and 2 slots
assign the 2 remaining pool devices, leave the filesystem set to auto
start the array to import the pool

Okay, so I've managed to kill the server entirely, so that's lovely.  Once again doing things you didn't ask for.

"Remove Pool" was grayed out while everything was running.  Attempted to stop the array again and it hung on unmounting disks although everything appeared to be spun down already it just couldn't complete unmounting.  So I tried rebooting and now everything is hosed.

The console says it is unable to connect to localhost.  Have I killed my system, and if so is there any way to recover the array where my actual data resides?  
 

8 minutes ago, Just Bill said:

So I tried rebooting and now everything is hosed.

Likely there is a problem with the pool cfg file, you should still be able to get the diagnostics from the console, post those.

I have absolutely no idea how to produce the diagnostics from the terminal, or where I'd place them to where I could then post them.  With the GUI down, I am beyond useless in linux. 

Right now I can load "Midnight Commander" but it doesn't seem to have access to anything other than itself.

From the terminal you use the ‘diagnostics’ command (as mentioned in the lin) and when it runs it tells you the result is put into the ‘logs’ folder on the flash drive.

I was able to create the diagnostics but I can't find the USB in Midnight Commander.  Where/what am I looking for.

NVM, found it.  

tower-diagnostics-20250505-1439.zip

Edit /config/pools/cache.cfg on the flash drive and change diskFsWidth="3" to "2", then reboot.

Brilliant!  Back up and running with the two disks in the pool.  It kicked off a parity check so it'll be at least 16 or so hours before I screw it all up again... LOL.  Thanks for the wild ride, it was very educational.  

Tomorrow after the parity check I may try and add a new parity disk and see what I can break then.  LOL