I don't know when it got to this point, but for some reason my cache drive thinks that it is a 2 device pool, even though I've only ever had one device there. I do not know if I misclicked at some point or what, but now anytime I restart my server, a btrfs balance starts and never goes anywhere (it will just get stuck at 0 out of 25)

Overall:
    Device size:		 715.41GiB
    Device allocated:		 293.12GiB
    Device unallocated:		 422.29GiB
    Device missing:		 238.47GiB
    Device slack:		     0.00B
    Used:			 250.63GiB
    Free (estimated):		 420.88GiB	(min: 243.81GiB)
    Free (statfs, df):		 121.30GiB
    Data ratio:			      1.09
    Metadata ratio:		      2.00
    Global reserve:		 345.55MiB	(used: 0.00B)
    Multiple profiles:		       yes	(data, metadata, system)

             Data      Data     Metadata Metadata  System   System                              
Id Path      single    RAID1    RAID1    DUP       RAID1    DUP      Unallocated Total     Slack
-- --------- --------- -------- -------- --------- -------- -------- ----------- --------- -----
 1 /dev/sdd1 239.00GiB 23.00GiB  1.00GiB   6.00GiB 32.00MiB 64.00MiB   207.85GiB 476.94GiB     -
 2 missing           - 23.00GiB  1.00GiB         - 32.00MiB        -   214.44GiB 238.47GiB     -
-- --------- --------- -------- -------- --------- -------- -------- ----------- --------- -----
   Total     239.00GiB 23.00GiB  1.00GiB   3.00GiB 32.00MiB 32.00MiB   422.29GiB 715.41GiB 0.00B
   Used      206.30GiB 21.58GiB 16.19MiB 582.61MiB    0.00B 64.00KiB                     

I have to cancel the balance anytime I restart.

How do I get rid of this detected second device that isn't actually part of the pool? I do have it only set to be 1 slot. I have taken the array down, unassigned the existing cache device, started the array again, chosen to "start array even with missing device", then restarted the server entirely, and reversed all of that and it still thinks there's a second slot.

root@Venus:~# btrfs fi usage /mnt/cache
Overall:
    Device size:                 715.41GiB
    Device allocated:            292.12GiB
    Device unallocated:          423.29GiB
    Device missing:              238.47GiB
    Device slack:                    0.00B
    Used:                        249.95GiB
    Free (estimated):            421.36GiB      (min: 244.00GiB)
    Free (statfs, df):           122.02GiB
    Data ratio:                       1.09
    Metadata ratio:                   2.00
    Global reserve:              343.09MiB      (used: 0.00B)
    Multiple profiles:                 yes      (data, metadata, system)

Data,single: Size:238.00GiB, Used:205.51GiB (86.35%)
   /dev/sdc1     238.00GiB

Data,RAID1: Size:23.00GiB, Used:21.64GiB (94.09%)
   /dev/sdc1      23.00GiB
   missing        23.00GiB

Metadata,RAID1: Size:1.00GiB, Used:16.19MiB (1.58%)
   /dev/sdc1       1.00GiB
   missing         1.00GiB

Metadata,DUP: Size:3.00GiB, Used:573.62MiB (18.67%)
   /dev/sdc1       6.00GiB

System,RAID1: Size:32.00MiB, Used:0.00B (0.00%)
   /dev/sdc1      32.00MiB
   missing        32.00MiB

System,DUP: Size:32.00MiB, Used:64.00KiB (0.20%)
   /dev/sdc1      64.00MiB

Unallocated:
   /dev/sdc1     208.85GiB
   missing       214.44GiB

With the array running type:

btrfs balance start -f -mconvert=dup,soft -dconvert=single,soft /mnt/cache

When that's done, type

btrfs device remove missing /mnt/cache

Then reimport the pool with the remaining device

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 1 slot

assign the pool device (sdc), leave the filesystem set to auto

start the array to import the pool

The pool should now be a single device pool; you can leave it like that or stop the array to add the other device to mirror the pool