Bug(?) in raid10 (raid10,f2 - lvm2 - xen)

["Tomasz Nowak" <tnowak-p@xxxxxxxxxxxxx>]

This bug comes ouf if you:

1. create a raid10,f2 device (ie /dev/md1)
2. set it as pv for vg
3. set a lv on that vg
4. export this /dev/vg/lv to xen domU as "disk" (ie xvdb)
  (that is the recommended way, feature of exporting
  devices as partitions will be dropped in the future
  probably)

As soon as domU get its hands on /dev/xvdb creating
a partition (xvdb1) there's a lot of filesystem
related errors, such as:
[...]
kernel: Buffer I/O error on device xvdb1, logical block 22
kernel: Buffer I/O error on device xvdb1, logical block 23
kernel: Buffer I/O error on device xvdb1, logical block 24
[...]

domU reports then:
[...]
raid10_make_request bug: can't convert block across chunks or bigger than 
256k 41952765 4
raid10_make_request bug: can't convert block across chunks or bigger than 
256k 41954301 4
raid10_make_request bug: can't convert block across chunks or bigger than 
256k 41954813 4
raid10_make_request bug: can't convert block across chunks or bigger than 
256k 41955325 4
[...]

Raid10,f2 is great (+80% read perfomance bust on my 2 disks only),
but unfortunably unusable as lvm backed xen device for the time
being.

PS. Centos 5.1 if that matters.

--
Tomek


--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html