March 29, 20242 yr Hi, I tried to remove the 2nd SSD of my cachepool. It used to be a raid1 - since I wanted to remove one disk, I converted it back to single. Then I removed the 2nd SSD from the config (new config). But I guess I missed a step. Now I can only wipe the data - which I want to avoid. So how do I either finish the conversion or restore it back? btrfs-select-super -s 1 /dev/nvme0n1p1 warning, device 2 is missing warning, device 2 is missing ERROR: cannot read chunk root ERROR: open ctree failed btrfs-select-super -s 1 /dev/nvme1n1 No valid Btrfs found on /dev/nvme1n1 ERROR: open ctree failed
March 29, 20242 yr Community Expert 3 minutes ago, patrickstigler said: since I wanted to remove one disk, I converted it back to single. Single in btrfs is between RAID0 and spanning, i.e. one big "drive" made of both, so both are required. Should have removed one while in RAID1 (only config that actually tolerates a drive loss), and THEN converted afterwards. Can you still add the 2nd drive back? If so it should just work again. Edited March 29, 20242 yr by Kilrah
March 29, 20242 yr Author warning, device 2 is missing warning, device 2 is missing ERROR: cannot read chunk root Label: none uuid: 53b61841-b900-45f6-a9b6-7446d79b46bc Total devices 2 FS bytes used 282.78GiB devid 1 size 465.76GiB used 14.00GiB path /dev/nvme0n1p1 *** Some devices missing
March 29, 20242 yr Community Expert Type sfdisk /dev/nvme1n1 then type 2048 and enter, post the results
March 29, 20242 yr Author Created a new DOS disklabel with disk identifier 0x8744012d. Created a new partition 1 of type 'Linux' and of size 931.5 GiB. Partition #1 contains a btrfs signature. Do you want to remove the signature? [Y]es/[N]o:
March 29, 20242 yr Community Expert Type N and enter to keep the signature, then type write followed by enter, post new output of btrfs fi show
March 29, 20242 yr Author Label: none uuid: 53b61841-b900-45f6-a9b6-7446d79b46bc Total devices 2 FS bytes used 282.78GiB devid 1 size 465.76GiB used 14.00GiB path /dev/nvme0n1p1 devid 2 size 931.51GiB used 290.06GiB path /dev/nvme1n1p1 looks better I guess?!
March 29, 20242 yr Community Expert Solution If OK then convert the pool to raid1, once done you can remove the device and the pool will automatically be converted to single.
April 1, 20242 yr Author Welcome :-D Unfortunately I still have some trouble, while I want to convert back to singe drive. root@Server:~# btrfs balance start -f -dconvert=single -mconvert=single /mnt/cache ERROR: error during balancing '/mnt/cache': Input/output error There may be more info in syslog - try dmesg | tail root@Server:~# dmesg | tail [276079.589839] BTRFS info (device nvme0n1p1): relocating block group 4693758574592 flags system|dup [276079.611808] BTRFS info (device nvme0n1p1): found 3 extents, stage: move data extents [276079.624923] BTRFS info (device nvme0n1p1): relocating block group 4692684832768 flags metadata|dup [276084.175103] BTRFS info (device nvme0n1p1): found 25823 extents, stage: move data extents [276086.154762] BTRFS info (device nvme0n1p1): relocating block group 1636480057344 flags data|raid1 [276086.348070] BTRFS warning (device nvme0n1p1): csum failed root -9 ino 922 off 130379776 csum 0x3019a12b expected csum 0x3ed50501 mirror 2 [276086.348083] BTRFS error (device nvme0n1p1): bdev /dev/nvme0n1p1 errs: wr 0, rd 0, flush 0, corrupt 807, gen 0 [276086.357072] BTRFS warning (device nvme0n1p1): csum failed root -9 ino 922 off 130379776 csum 0x3019a12b expected csum 0x3ed50501 mirror 2 [276086.357079] BTRFS error (device nvme0n1p1): bdev /dev/nvme0n1p1 errs: wr 0, rd 0, flush 0, corrupt 808, gen 0 [276086.604873] BTRFS info (device nvme0n1p1): balance: ended with status: -5
April 1, 20242 yr Community Expert Balance is aborting because it's finding some corrupt data, if you don't have ECC RAM it may be a good idea to run memtest first, then scrub the pool and look at the syslog for the list of corrupt files, delete/restore those from a backup, then run another scrub to confirm there are no more errors and try to convert again.
April 1, 20242 yr Author I ran Memtest, the whole weekend, last week :-D I got 2 broken modules. So the other ones, currently working fine - at leased I'd say so. I tried to scrub it. But there is one Uncorrectable error. So what to do now? UUID: 53b61841-b900-45f6-a9b6-7446d79b46bc Scrub started: Mon Apr 1 16:58:22 2024 Status: aborted Duration: 0:07:20 Total to scrub: 354.71GiB Rate: 795.44MiB/s Error summary: csum=1 Corrected: 0 Uncorrectable: 1 Unverified: 0
April 1, 20242 yr Community Expert 2 hours ago, JorgeB said: look at the syslog for the list of corrupt files, delete/restore those from a backup, then run another scrub to confirm there are no more errors and try to convert again.
April 1, 20242 yr Author only broken sectors ... I justed wiped the disk, and recovered from the backup. Took a while, but it seems to work.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.