Re: Many I/O errors with btrfs on MicroSD and RPi

[Mobin Aydinfar <mobin@mobintestserver.ir>]

On ۱۴۰۳/۶/۱۹ ۱:۴۰, Qu Wenruo wrote:
> 
> 
> 在 2024/9/9 07:00, Mobin Aydinfar 写道:
>> Hi.
>>
>> My Raspberry Pi is down because btrfs reports I/O error. So I removed my
>> MicroSD from RPi and I mounted it on my computer with ram-reader. btrfs
>> check reports many problems.
>>
>> The documention says don't use btrfs check --repair unless you know what
>> are you doing so I didn't use it.
> 
> That's the correct way to go.
> 

:)
> Unfortunately from the output, it looks quite some of your metadata is
> full of garbage now:
> 
>   BTRFS warning (device sdc2): tree block 121929728 mirror 1 has bad
> bytenr, has 15771731761493389218 want 121929728
>   BTRFS warning (device sdc2): tree block 121929728 mirror 2 has bad
> bytenr, has 17159496520340310444 want 121929728
> 
> The "bytenr" is a u64 representing the logical address, should at least
> always be aligned to sector size (4K in most cases).
> 
> Furthermore it shows both copy is corrupted, with different corruption.
> 
> It looks like the sdcard itself is not reliable thus a big chunk of data
> is either discarded or filled with garbage.
> 
> So far no transid error, thus at least its FLUSH/FUA command handling
> seems sane.

Looks like my MicroSD just collapsed because I see some fatal errors 
even on my FAT32 boot device.

I replaced the failing MicroSD. The copied data was enough for getting 
the system up and running again. Thanks for your detailed explanation 
and your great work on btrfs :)

Best Regards.

> Thanks,
> Qu
>>
>> I mounted the MicroSD and I copied all of its readable content and I
>> tried to use "btrfs scrub start" but it got aborted after ~10GiB.
>>
>> I check RPi's dmesg everyday to ensure there is no under voltage, I'm
>> sure that there is no under voltage issue but I had multiple power
>> outages in 3 last months.
>>
>> I don't have dmesg from RPi, but I have dmesg from my computer (See the
>> attachment).
>>
>> btrfs scrub status:
>> UUID:             4c3054bf-5704-4afe-be0d-d2300359d493
>> Scrub resumed:    Mon Sep  9 00:34:28 2024
>> Status:           aborted
>> Duration:         0:11:00
>> Total to scrub:   33.99GiB
>> Rate:             17.01MiB/s
>> Error summary:    csum=52
>>    Corrected:      4
>>    Uncorrectable:  48
>>    Unverified:     0
>>
>> My computer uname:
>> Linux Mobin-Mehdi-PC 6.10.7-0-generic #1 SMP PREEMPT_DYNAMIC Thu Aug 29
>> 16:26:29 UTC 2024 x86_64
>>
>> btrfs --version:
>> btrfs-progs v6.10.1
>> -EXPERIMENTAL -INJECT -STATIC +LZO +ZSTD +UDEV +FSVERITY +ZONED
>> CRYPTO=builtin
>>
>> btrfs fi show:
>> Label: none  uuid: 4c3054bf-5704-4afe-be0d-d2300359d493
>>      Total devices 1 FS bytes used 33.79GiB
>>      devid    1 size 59.02GiB used 39.07GiB path /dev/sdc2
>>
>> btrfs fi df /mnt/Removable:
>> Data, single: total=35.01GiB, used=33.59GiB
>> System, DUP: total=32.00MiB, used=16.00KiB
>> Metadata, DUP: total=2.00GiB, used=200.53MiB
>> GlobalReserve, single: total=54.44MiB, used=0.00B
>>
>> dmesg: As mentioned, Check the attachment.
>>
>> Best Regards.