Re: Write time tree block corruption detected

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

On 2021/6/9 上午6:39, Qu Wenruo wrote:
>
>
> On 2021/6/8 下午10:45, auxsvr wrote:
>> On Monday, 7 June 2021 14:55:10 EEST you wrote:
>>> It seems that every time free space is less than 1 GiB and I use
>>> zypper, I get a similar corruption message.
>>
>> Correction: I just got:
>>
>> [26355.330817] BTRFS critical (device sda2): corrupt leaf: root=3
>> block=537007243264 slot=0 devid=1 invalid bytes used: have 64503152640
>> expect [0, 64424509440]
>
> The line itself should explain itself, we're over allocating the device.
>
>> [26355.330819] BTRFS error (device sda2): block=537007243264 write
>> time tree block corruption detected
>
> Mind to dump the device tree?
>
> You can dump it with the following command, and the output contains
> nothing confidential, but only device extent layout:
>
> # btrfs ins dump-tree -t dev /dev/sda2

Sorry, you may also need to dump the super block:

# btrfs ins dump-super /dev/sda2

One of my concern is, the device used bytes may differ between super
block and device tree.

Thanks,
Qu
>
>>
>> and:
>>
>> # btrfs fi df /
>> Data, single: total=55.98GiB, used=43.04GiB
>> System, single: total=32.00MiB, used=16.00KiB
>> Metadata, single: total=2.00GiB, used=1.17GiB
>> GlobalReserve, single: total=512.00MiB, used=0.00B
>>
>> # btrfs fi show /
>> Label: none  uuid: 44c67fa4-e2c4-4da5-9d07-98959ff77bc4
>>          Total devices 1 FS bytes used 44.22GiB
>>          devid    1 size 60.00GiB used 60.07GiB path /dev/sda2
>
> My current guess is, there is something wrong in the past, leaving the
> accounting wrong.
>
> It leaves some ghost used bytes, thus causing the problem.
>
> Your previous btrfs-check is able to detect that, but unfortunately we
> don't have the ability to repair it in btrfs-check right now.
>
> I'll work on repair ability soon, thus if you can keep the fs, there is
> a high chance to repair it.
>
> Thanks,
> Qu
>
>>
>> 60.07GiB used again!
>>
>> Regards,
>> Petros
>>
>>
>>
>>