Created attachment 184691 [details]
dmesg output

Below is what I posted to the btrfs mailing list.  I was asked create a bug report and attach my entire dmesg output, so I'm doing that here.


I have a four-drive RAID1 array, and since yesterday, some problem has
rendered it unmountable (read/write anyhow).  One drive reports a read
error, so maybe the drive is failing, but I've had that happen before,
and it was easy to swap in a new drive.  This time, two more drives
are reporting that they "failed to read the system array."  I managed
to mount it read-only (by specifying the node of the fourth drive) and
rsync everything to a backup drive.  Now I'd like to try to repair.
This is where I'm running into problems.  Since I can't mount it
read-write, I can't do a scrub, so I tried "btrfs check --repair", and
this is what I got:

# btrfs check --repair /dev/sde
enabling repair mode
Checking filesystem on /dev/sde
UUID: ecdff84d-b4a2-4286-a1c1-cd7e5396901c
checking extents
ref mismatch on [1667931533312 524288] extent item 1, found 2
attempting to repair backref discrepency for bytenr 1667931533312
Ref doesn't match the record start and is compressed, please take a
btrfs-image of this file system and send it to a btrfs developer so
they can complete this functionality for bytenr 1667931639808
failed to repair damaged filesystem, aborting

Since this specifically told me to contact a developer, I figured this
is something you guys want to know about.  :)

Also, I was wondering if perhaps someone can help me figure out how to
repair it.

There are only two files that appear to be unrecoverable when I rsync,
and I can restore those from an earlier backup.  Since I can't mount
read/write, I can't go and delete those files, so I seem to be stuck.