Cache Pool btrfs Raid1 unmountable

[Schulmeister]

Hi all,

I have a Problem - as usual 🙂

I have a Cachepool with two HDDs running on btrfs, raid1.

Usually /mnt/data

 

After an error in the second disk I tried to replace the disk - now everything ist unmountable

 

I tried:

 

root@RD6:~# btrfs rescue super-recover -v /dev/sdc1
All Devices:
        Device: id = 1, name = /dev/sdc1

Before Recovering:
        [All good supers]:
                device name = /dev/sdc1
                superblock bytenr = 65536

                device name = /dev/sdc1
                superblock bytenr = 67108864

                device name = /dev/sdc1
                superblock bytenr = 274877906944

        [All bad supers]:

All supers are valid, no need to recover

 

----------------

 

root@RD6:~# btrfs check --readonly --force /dev/sdc1
Opening filesystem to check...
bad tree block 8152159748096, bytenr mismatch, want=8152159748096, have=0
Couldn't read tree root
ERROR: cannot open file system

 

--------------------

 

btrfs restore --dry-run -d /dev/sdc1 /mnt/disk7/restore/data/
bad tree block 8152159748096, bytenr mismatch, want=8152159748096, have=0
Couldn't read tree root
Could not open root, trying backup super
bad tree block 8152159748096, bytenr mismatch, want=8152159748096, have=0
Couldn't read tree root
Could not open root, trying backup super
bad tree block 8152159748096, bytenr mismatch, want=8152159748096, have=0
Couldn't read tree root
Could not open root, trying backup super

 

 

Nothing seems to work.

 

that beeing my third time a "so called" raid filesystems (ZFS and btrfs) failed me I am very disappointed with these things.

What is the sense of spending double the money on a security that does not do anything

 

I shall buy myself a NAS and backup everything everyday and stop bothering with btrfs or zfs crap.

 

Thank you all in advance for your time and support.

rd6-diagnostics-20240117-1028.zip

Edited January 17 by Schulmeister
forgot to thanks everybody

[JorgeB]

I'm trying to follow the sequence of events, from what I see, before removing the pool device the metadata was balance to single:

 

Jan 17 01:32:20 RD6 kernel: BTRFS info (device sdc1): balance: start -f -mconvert=single -sconvert=single

 

Then you removed the device and the pool not surprisingly failed to mount, do you remember why it was balanced to single?  Is the old pool disk ST16000VN001-2RV103_ZL22EVWC still intact?

 

P.S. btrfs is detecting data corruption in multiple pools, xfs also detecting metadata corruption, suggesting you may have an underlying issue:

 

Jan  6 01:20:47 RD6 kernel: XFS (md1p1): Metadata corruption detected at xfs_dinode_verify+0xa0/0x732 [xfs], inode 0x55c36e0b2 dinode
Jan  6 01:20:47 RD6 kernel: XFS (md1p1): Unmount and run xfs_repair

 

[Schulmeister]

Hi Jorge,

I tried to change the raid1 to single to remove the second drive - well that seemed to cause a problem.

It is not that bad as far as there's not very important data on it.

The second drive ZL22EVWC is still intact and connected.

 

The PS you wrote is very disturbing:

P.S. btrfs is detecting data corruption in multiple pools, xfs also detecting metadata corruption, suggesting you may have an underlying issue:

 

What should i do ?

 

[itimpi]

45 minutes ago, Schulmeister said:

tried to change the raid1 to single to remove the second drive - well that seemed to cause a problem.

It definitely would as it stops you being able to remove a drive keeping data intact.

[Schulmeister]

26 minutes ago, itimpi said:

It definitely would as it stops you being able to remove a drive keeping data intact.

Is there any way to get to the Data ?

[JorgeB]

If the old disk is still intact you can try typing with the array stopped:

 

sgdisk -o -a 8 -n 1:32K:0 /dev/sdX

 

Replace X with correct letter, then post the output of:

 

btrfs fi show

 

[Schulmeister]

I will try this tonite, thanks

[Schulmeister]

root@RD6:~# sgdisk -o -a 8 -n 1:32K:0 /dev/sdi
Creating new GPT entries in memory.
The operation has completed successfully.
root@RD6:~# btrfs fi show
Label: none  uuid: 431bb516-355f-42bf-9966-f4413b80ad33
        Total devices 2 FS bytes used 1.21TiB
        devid    1 size 3.64TiB used 1.26TiB path /dev/nvme0n1p1
        devid    2 size 3.64TiB used 1.26TiB path /dev/nvme1n1p1

Label: none  uuid: c9259cf9-d19a-47b4-987e-42b0f5f82617
        Total devices 2 FS bytes used 4.82TiB
        devid    1 size 14.55TiB used 6.22TiB path /dev/sdc1
        devid    2 size 14.55TiB used 6.22TiB path /dev/sdi1

Label: none  uuid: e474e21e-688b-4eda-8e90-9177c495d366
        Total devices 1 FS bytes used 12.08GiB
        devid    1 size 894.25GiB used 16.02GiB path /dev/sdh1
 

[Schulmeister]

20 hours ago, JorgeB said:

If the old disk is still intact you can try typing with the array stopped:

 

sgdisk -o -a 8 -n 1:32K:0 /dev/sdX

 

Replace X with correct letter, then post the output of:

 

btrfs fi show

 

That did it - I think.

I did as you told me, took every HDD out of that Cachepool (deleting the Cachepool) and put the HDDs in Unassigned Devices - thankfully all files were readable and I copied everything to the array.
I had this Cachepool of two HDDs (16TB each) only to speed up read/write operations - it is not the way it's intended, i know.

Is there a way to have a Disk/Raid/Pool that can speed up things ?

 

Anyway, as always I am very grateful for the help you provide here - it really makes unraid stand out.

 

[JorgeB]

1 hour ago, Schulmeister said:

I had this Cachepool of two HDDs (16TB each) only to speed up read/write operations - it is not the way it's intended, i know.

Is there a way to have a Disk/Raid/Pool that can speed up things ?

Raid1 won't speed up writes, it would speed up reads a little if you use zfs, raid0 would speed up reads and writes, but there's no redundancy.