I swapped a failing cache drive (sdg) with a new same-size drive (sdd). BTRFS balance still says it found 'extents' on the bad drive

[Blaze9]

Hey ya'll

 

I have a 3 drive cache pool in raid1. All 3 drives are 1TB, I have 1.5TB of usable space.

One of my drives (sdg) has given me errors, so I replaced the drive with an identical new ssd (sdd). 

 

sdg is still connected as an unassigned drive.

 

When I do a btrfs balance, my logs are still showing entries saying

 

Feb 16 02:24:21 Mk4Alpha kernel: BTRFS info (device sdg1): found 14 extents
Feb 16 02:24:22 Mk4Alpha kernel: BTRFS info (device sdg1): relocating block group 10354819727360 flags data|raid1

 

I'm not sure why? I don't have access right now to the server to disconnect the failing drive. Is that the only way to truly remove the drive? If I bring the array down, is Wireguard still active? That's my only way into the server right now.

 

I tried doing 

btrfs device delete missing /mnt/cache

but this says `ERROR: error removing device 'missing': no missing devices found to remove`

And the following is my output for the pool from btrfs

 

root@Mk4Alpha:~# btrfs filesystem show
Label: none  uuid: 7ed018b8-03ec-408f-b72f-64210d53b1ed
        Total devices 3 FS bytes used 540.46GiB
        devid    2 size 931.51GiB used 370.00GiB path /dev/sdf1
        devid    3 size 931.51GiB used 368.03GiB path /dev/sdd1
        devid    4 size 931.51GiB used 368.03GiB path /dev/sdc1

Thanks!

mk4alpha-diagnostics-20220216-0230.zip

Edited February 16, 2022 by Blaze9

[JorgeB]

Device sdg was successfully removed from the pool, any mentions of it should stop once you reboot, possibly after an array re-start.