BTRFS errors & read-only after restoring system partition

English Install/Boot/Login
1

After restoring my system partition with clonezilla when I boot I get the partition forced read-only due to BTRFS errors. I’m quite sure the clonezilla partition image is OK

dmesg output this errors:

[   79.693242] systemd-journald[594]: File /var/log/journal/70cf9b43b18548dba2f9e8d59e694030/user-1000.journal corrupted or uncleanly shut down, renaming and replacing.
[  102.571835] BTRFS error (device nvme0n1p5): level verify failed on logical 21189033984 mirror 1 wanted 1 found 0
[  102.572056] BTRFS error (device nvme0n1p5): level verify failed on logical 21189033984 mirror 2 wanted 1 found 0
[  102.572152] BTRFS warning (device nvme0n1p5): Skipping commit of aborted transaction.
[  102.572155] BTRFS error (device nvme0n1p5: state A): Transaction aborted (error -5)
[  102.572157] BTRFS: error (device nvme0n1p5: state A) in cleanup_transaction:2021: errno=-5 IO failure
[  102.572160] BTRFS info (device nvme0n1p5: state EA): forced readonly
[  143.474666] systemd-journald[594]: /var/log/journal/70cf9b43b18548dba2f9e8d59e694030/user-1000.journal: Journal file corrupted, rotating.
[  143.474681] systemd-journald[594]: Failed to rotate /var/log/journal/70cf9b43b18548dba2f9e8d59e694030/system.journal: Read-only file system
[  143.474687] systemd-journald[594]: Failed to rotate /var/log/journal/70cf9b43b18548dba2f9e8d59e694030/user-1000.journal: Read-only file system
[  143.475018] systemd-journald[594]: Failed to write entry to /var/log/journal/70cf9b43b18548dba2f9e8d59e694030/user-1000.journal (30 items, 889 bytes) despite vacuuming, ignoring: Bad message

Can this errors came from restoring the partition?

Looks like hardware problem but I have Windows instaled on another partition of the same disk and it works OK. Is an SSD disk.

I restored again the partition with same result, clonezilla gives no errors and say image is OK.

I try to fix using gparted live, gparted say it can not fix the partition, I try “btrfs check --force /dev/nvme0n1p5”, it fix tons of errors and finish unexpectedly.

2

That is typical bit flip most commonly associated with bad memory. Run memory check for extended period (24 hours at least).

1 Like
3

I had those kind of errors plenty of times after restoring BTRFS filesystems in the past, using Clonezilla. That’s exactly the reason why I don’t use it with BTRFS volumes anymore.

1 Like
4

I´ve run memtest 8 hour without errors.
Formating the partition before restoring could help?

5

Sure. Windows doesn’t bother.

https://btrfs.readthedocs.io/en/latest/Hardware.html#hardware-as-the-main-source-of-filesystem-corruptions

Formatting won’t fix the root cause.

1 Like
6

You mean windows doesn’t check it and opensuse does?

7

It doesn’t: Comparison of file systems - Wikipedia

8

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.