Open gwemmie opened 1 year ago
This error has finally just now occurred on ext4... So I guess it's not exclusive to btrfs. Googling it now, without looking for just btrfs hits, it seems to be widespread across rock64 boards. Maybe even a hardware issue? I haven't finished researching.
Hit this problem on oDroid C2, Armbian Bookwork (happened on bullseye also).
Got a mix of disks and filesystems attached, with things running on top.
Samba, NFS and GlusterFS.
I saw the hang before I installed GlusterFS, so I don't believe that is involved, but the rest was present.
The hints I got from a terminal that was open before the hand happened:
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.467099] Internal error: Oops: 96000005 [#1] PREEMPT SMP
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.647127] Process .NET ThreadPool (pid: 489238, stack limit = 0x00000000bf005c5e)
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.730445] Code: a903e7b8 b9402262 b9401020 11000400 (f8626ad8)
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.467099] Internal error: Oops: 96000005 [#1] PREEMPT SMP
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.647127] Process .NET ThreadPool (pid: 489238, stack limit = 0x00000000bf005c5e)
Message from syslogd@files at Apr 7 10:11:19 ...
kernel:[213832.730445] Code: a903e7b8 b9402262 b9401020 11000400 (f8626ad8)
Point being, this does not seem to be isolated to just one platform/OS
On Armbian Buster, Rock64 (RK3328). This was also happening on the latest backported version of 5.10, I think 5.10.72? I tried upgrading to fix it to no avail.
I have a couple 8TB HDDs that I use as backups. All I did was reformat one of them as btrfs and start copying the data back over from the other (still EXT4). I tried rsync and simple cp. Without fail, a few minutes into the operation, the copy operation will hang indefinitely, and all open terminals receive these kernel messages:
After that, every subsequent copy operation from the EXT4 drive to this new btrfs partition will hang until I reboot the rock64.
However, the drive otherwise continues to work fine. fsck and btrfs scrub find no problems. The super weird thing is I can still write files, and copy a small shell script from the home folder, all just fine--but copying from the other 8TB drive will immediately hang. I have had zero issues with either drive until trying btrfs now. I'm able to rsync between them just fine when they're both EXT4.
journalctl.txt specific-error.txt