On Thu, 14 Mar 2013 22:04:17 -0600
Chris Murphy [off-list ref] wrote:

2. sdg is encountering UNC read errors, probably bad sectors, why it's new right?

It's not new, in fact it is in a questionable condition, but that's not the
issue I am looking for help with, in fact I am glad if me using this disk
suddenly helps uncover/solve some bug in mdadm :) I doubt it receives too much
testing with regard to its behavior in such rare circumstances (encountering
bad sectors on a member configured as write-mostly with btrfs on top of
everything, etc).

3. md3 might be getting the mirrored copy off sdf, I can't actually tell

[  771.438945] btrfs: bdev /dev/md3 errs: wr 0, rd 9, flush 0, corrupt 0, gen 0

That line is btrfs reporting read errors from md3, not corruptions.

[  771.439011] ata11: EH complete
[  771.439035] btrfs no csum found for inode 31016 start 5668864
[  771.479594] btrfs csum failed ino 31016 off 5668864 csum 2731683453 private 0

What are your btrfs mount options? Are you using nodatasum or nodatacow?

rw,noatime,nodiratime,compress=lzo,ssd,discard,space_cache,inode_cache

-- 
With respect,
Roman