Trying to tackle https://github.com/kdave/btrfs-progs/issues/541

Trying to tackle https://github.com/kdave/btrfs-progs/issues/541

[Raúl Sánchez Siles]

[-- Attachment #1: Type: text/plain, Size: 2013 bytes --]

Hello!

After a mount error "parent transid verify failed" I went to offline check the 
partition and I came accross this issue. After this, I backup all data I 
could get from the partition (via btrs restore). As a sidenote I also tried 
mounting with the rescue=usebackuproot with the same outcome.

Currently I have a reproducible scenario to look into the issue and provided 
lack of activity in the issue I thought I may be of help to provide more 
information so issue can be fixed. The btrfs consistency issue looks serious 
so even in case of successful debugging I don't expect to recover access to 
that partition.

OS and SW:

- Debian sid
- btrfs-progs 6.17
- linux 6.16.11

Mount error:
[  120.652356] BTRFS: device fsid 06dc5f24-a16f-4520-993a-xxxxxxxxxxxx devid 
1 transid 968758 /dev/sda3 (8:3) scanned by pool-2 (4449)
[  120.653973] BTRFS info (device sda3): first mount of filesystem 06dc5f24-
a16f-4520-993a-xxxxxxxxxxxx
[  120.653994] BTRFS info (device sda3): using crc32c (crc32c-x86) checksum 
algorithm
[  123.225415] BTRFS info (device sda3): start tree-log replay
[  123.523786] BTRFS error (device sda3): parent transid verify failed on 
logical 229603573760 mirror 1 wanted 968173 found 966625
[  123.538334] BTRFS error (device sda3): parent transid verify failed on 
logical 229603573760 mirror 2 wanted 968173 found 966625
[  123.538372] BTRFS error (device sda3): failed to run delayed ref for 
logical 487464960 num_bytes 4096 type 184 action 1 ref_mod 1: -5
[  123.538383] BTRFS error (device sda3 state A): Transaction aborted (error 
-5)
[  123.538387] BTRFS: error (device sda3 state A) in 
btrfs_run_delayed_refs:2159: errno=-5 IO failure
[  123.538395] BTRFS: error (device sda3 state EA) in btrfs_replay_log:2094: 
errno=-5 IO failure (Failed to recover log tree)
[  123.544693] BTRFS error (device sda3 state EA): open_ctree failed: -5

Please CC me as I'm not subscribed to the list.

  Best Regards,
-- 
Raúl Sánchez Siles

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: Trying to tackle https://github.com/kdave/btrfs-progs/issues/541

[Qu Wenruo]

在 2025/10/10 19:25, Raúl Sánchez Siles 写道:
> Hello!
> 
> After a mount error "parent transid verify failed" I went to offline check the
> partition and I came accross this issue. After this, I backup all data I
> could get from the partition (via btrs restore). As a sidenote I also tried
> mounting with the rescue=usebackuproot with the same outcome.
> 
> Currently I have a reproducible scenario to look into the issue and provided
> lack of activity in the issue I thought I may be of help to provide more
> information so issue can be fixed. The btrfs consistency issue looks serious
> so even in case of successful debugging I don't expect to recover access to
> that partition.
> 
> OS and SW:
> 
> - Debian sid
> - btrfs-progs 6.17
> - linux 6.16.11
> 
> Mount error:
> [  120.652356] BTRFS: device fsid 06dc5f24-a16f-4520-993a-xxxxxxxxxxxx devid
> 1 transid 968758 /dev/sda3 (8:3) scanned by pool-2 (4449)
> [  120.653973] BTRFS info (device sda3): first mount of filesystem 06dc5f24-
> a16f-4520-993a-xxxxxxxxxxxx
> [  120.653994] BTRFS info (device sda3): using crc32c (crc32c-x86) checksum
> algorithm
> [  123.225415] BTRFS info (device sda3): start tree-log replay
> [  123.523786] BTRFS error (device sda3): parent transid verify failed on
> logical 229603573760 mirror 1 wanted 968173 found 966625
> [  123.538334] BTRFS error (device sda3): parent transid verify failed on
> logical 229603573760 mirror 2 wanted 968173 found 966625

Extent tree corruption, metadata COW is broken.

It looks like some writes are lost. Btrfs check should report the same 
problem.

It is not log tree causing the bug.

I'd recommend posting the device model and check if there is any known 
report with transid mismatch.

Sometimes you can not really just trust their firmwares.

Thanks,
Qu

> [  123.538372] BTRFS error (device sda3): failed to run delayed ref for
> logical 487464960 num_bytes 4096 type 184 action 1 ref_mod 1: -5
> [  123.538383] BTRFS error (device sda3 state A): Transaction aborted (error
> -5)
> [  123.538387] BTRFS: error (device sda3 state A) in
> btrfs_run_delayed_refs:2159: errno=-5 IO failure
> [  123.538395] BTRFS: error (device sda3 state EA) in btrfs_replay_log:2094:
> errno=-5 IO failure (Failed to recover log tree)
> [  123.544693] BTRFS error (device sda3 state EA): open_ctree failed: -5
> 
> Please CC me as I'm not subscribed to the list.
> 
>    Best Regards,

Re: Trying to tackle https://github.com/kdave/btrfs-progs/issues/541

[Raúl Sánchez Siles]

Hi!

Thanks for the reply.

El 10 de octubre de 2025 22:47:15 CEST, Qu Wenruo <wqu@suse.com> escribió:
>
>
>在 2025/10/10 19:25, Raúl Sánchez Siles 写道:
[...]
>> 
>> Currently I have a reproducible scenario to look into the issue and provided
>> lack of activity in the issue I thought I may be of help to provide more
>> information so issue can be fixed. The btrfs consistency issue looks serious
>> so even in case of successful debugging I don't expect to recover access to
>> that partition.
>> 
[...]
>> 
>> Mount error:
>> [  120.652356] BTRFS: device fsid 06dc5f24-a16f-4520-993a-xxxxxxxxxxxx devid
>> 1 transid 968758 /dev/sda3 (8:3) scanned by pool-2 (4449)
>> [  120.653973] BTRFS info (device sda3): first mount of filesystem 06dc5f24-
>> a16f-4520-993a-xxxxxxxxxxxx
>> [  120.653994] BTRFS info (device sda3): using crc32c (crc32c-x86) checksum
>> algorithm
>> [  123.225415] BTRFS info (device sda3): start tree-log replay
>> [  123.523786] BTRFS error (device sda3): parent transid verify failed on
>> logical 229603573760 mirror 1 wanted 968173 found 966625
>> [  123.538334] BTRFS error (device sda3): parent transid verify failed on
>> logical 229603573760 mirror 2 wanted 968173 found 966625
>
>Extent tree corruption, metadata COW is broken.
>
>It looks like some writes are lost. Btrfs check should report the same problem.
>

It is right that btrfs check gave the same report (I forgot to mention it)

>It is not log tree causing the bug.
>
>I'd recommend posting the device model and check if there is any known report with transid mismatch.
>

The HD model is Toshiba DTB420 2TB. Is there a specific place to look for transid mismatch reports or do you mean just a generic search?

>Sometimes you can not really just trust their firmwares.
>

I have been using this drive for at least a couple of years. There has been several power outages, but none of them lead to a such serious trouble. Actually I think when the FS stalled, the power supply was OK.

>Thanks,
>Qu
>

So still, let me know if I can be of help debugging/managing this kind of scenarios.

Best Regards,