[mythtv-users] OT Help: mdadm segfault

[Steve MacLaren]

Well, my Ubuntu Feisty backend has been down for about a week now following
a crash that sent it into an endless rebooting loop.  My system used a
single IDE drive (hda) for the operating system, and a 4-disk SATA raid5
array (sda/sdb/sdc/sdd) for my recordings and videos, mounted as
/var/lib/mythtv/

Using the Live CD, I was able to boot up and see that the /var/log/messages
file just before the crash had filled up with
"md: array md0 already has disks!"

I have since been able to boot up by selecting the alternate kernel in grub
(2.6.20-15 was installed alongside 2.6.20-16 with my Feisty installation).
In order to get the machine to boot, I also had to comment out the /dev/md0
line in /etc/fstab and rename /etc/mdadm.conf.  When I do this, I see the
following in /var/log/messages:

Sep 16 21:27:42 mediaserver kernel: [   34.647825] EXT3-fs: mounted
filesystem with ordered data mode.
Sep 16 21:27:42 mediaserver kernel: [   34.678639] md: md0 stopped.
Sep 16 21:27:42 mediaserver kernel: [   34.697939] mdadm[2693]: segfault at
0000000000000004 rip 000000000041567c rsp 00007fff97aaec20 error 4
Sep 16 21:27:42 mediaserver kernel: [   34.706476] md: md0 stopped.
Sep 16 21:27:42 mediaserver kernel: [   34.733561] md: hda has invalid sb,
not importing!
Sep 16 21:27:42 mediaserver kernel: [   34.733565] md: md_import_device
returned -22

Because /dev/md0 no loger exists, any attempts I have made to mount or
diagnose the status of the raid5 array have been futile.  Does anyone with
software raid experience have any suggestions on how I could go about
recovering the data on the array?  What would cause mdadm to segfault?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mythtv.org/pipermail/mythtv-users/attachments/20070917/3420a715/attachment.htm