check/qgroup-verify.c:546: account_all_refs: BUG_ON ref->num_bytes != num_bytes triggered, value 1
#358
Labels
Comments
|
Now i have done a btrfs check and repair with the new updated version: 5.11.1 It seems like the BUG is still there. The brtfs check doesn't complete it stops.. |
SoulInfernoDE
changed the title
ref->num_bytes != num_bytes triggered, value 1
|
The warning The checksum error means some tree blocks are zeroed out. I guess it's your quota root cleared but by somehow there is still an quota root item in your root tree? Please provide the following output: |
|
Okay, here is the output:
|
|
OK, got the reason. It looks like btrfs-progs can't handle half dropped trees, thus causing the problem. The tree involved is If your fs is still being able to be mounted RW, please mount it, run Then btrfs check should not trigger the qgroup problem. |
|
Wow your reply is as fast as lightning. I'll reboot to prod system and will check there and reply here afterwards.. |
|
|
You need to mount it, and execute the progs on the mount point, not the device. |
|
Oh sorry my fault 👍 I have done:
So the output of btrfs check is: It did indeed help. Only two lines left to fix instead of hundreds. Maybe you have an idea for that also? 👍 best regards |
|
It could be related to the --readonly, normally we don't recommand using that option, unless it's unavoidable. Thus still recommended to do a proper btrfs check on the unmounted fs. Anyway, this should not be a big thing, as qgroup shouldn't really cause any obvious impact. |
|
Okay i rebooted and tried to btrfs check again on unmounted /dev/sda6 I am not sure if Clonezilla will complete when this errors are seen, because at the end of the command it says: Abgebrochen which means in english something like: cancelled |
|
For the worst case, you can try to disable qgroup by Then you can check if the tool still works. As despite the qgroup tree block read error, there seems to be nothing wrong at all of your fs. |
|
Will do that tomorrow as i need to do some rl stuff now. I'll report back if it worked. 👍 |
|
It should be a btrfs-progs bug, firstly it didn't follow kernel dropped trees (this is kinda known). Will look into the problem soon. |
|
Hi @adam900710, I did a qgroup rescan /dev/sda6 which cleared the two errors mentioned above. I then tried to btrfs check /dev/sda6 which also cancelled the scan. But now i don't see any plausible error now.. Hmh ..will update my post as soon as i tried now to backup with qgroup disabled again after rescan. Also is it save to destroy my qgroups before disable in case to fix the abort? btrfs check /dev/sda6 with qgroup disabled went through the check at least |
|
So there are two bugs, and even maybe unrelated. For the clonezilla bug, I have no idea as I never use it. For the qgroup part, if your fs contains no confidential filenames, would you mind to provide the btrfs-image dump of it? Please note that, although this dump won't provide any data, but it contains file and directory names. |
I have the exact same problem. And, this error does occur with quota disabled. Disabling and then re-enabling quota brings back the issue. I have tried upto btrfs-progs v5.15.1 Is there any other fix that disabling quota? Because I use timeshift, and it works better with quota enabled. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello,
when i tried to backup my btrfs root partition with clonezilla it throws errors and then aborted.
I booted in a live linux live env. and installed btrfs-progs 5.11 and did a btrfs check --readonly /dev/sda6:
I really need to copy this partition to my new external ssd from my internal ssd.
note: the system behaves normal i didn't noticed any bad behaviour while using normal.
I just ran into this when i tried to copy the btrfs partition to the new ssd.
Linux live kernel version im using:
In my prod. laptop system i am using kernel 5.11.7
best regards
SI
The text was updated successfully, but these errors were encountered: