kernel-packages team mailing list archive

[Tapani Tarvainen <ubuntu@xxxxxxxxxxxxxxxxxxxxx>]

Comparing moved data against backup I found exactly one corrupt file
(out of about 1.4 million), despite thousands of those "bio too big"
messages. Also, I found one "buffer I/O error" in the logs. So I suspect
the "bio too big" messages probably did not cause data corruption at all
(even  though the earlier bug causing same message reportedly did). So
it would be less bad than I originally thought.

Also, I have been unable to reproduce the error with smaller dataset and
can't easily recreate the original situation or even one of almost same
size anymore, unless I buy more disks. I may do that ahead of schedule
just to retest this, but at the moment I can't test if mainline kernel
would've made a difference.

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1382335

Title:
  mdadm+lvm+cryptsetup+big disks: "bio too big device"

Status in “linux” package in Ubuntu:
  Incomplete

Bug description:
  I installed two new 6TB disks (WD Red) as RAID1 and pmove'd stuff over from old
  5x3TB RAID10 array to it (actually the move is still in progress, going to take some
  17 hours or so) and I get lots of these (4971 so far) in dmesg:

  [58693.807553] bio too big device dm-7 (664 > 8)
  [58694.037060] bio too big device dm-7 (1024 > 8)

  Looks like an old kernel bug that was supposedly fixed in 2012:
  https://lkml.org/lkml/2012/10/15/398

  The LV that apparently triggered that is 3TB in size (90% full) and
  encrypted with cryptsetup.

  I don't know if this is reproducible but I can try - it'd just take some
  two days every time (starting with restoring 3TB from backup).

  Ubuntu 14.04.1, 3.13.0-37-generic x86_64 kernel,
  Intel DQ57TM mobo, LSI 9211-8i SAS/SATA controller.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1382335/+subscriptions