On Fri, August 21, 2009 12:52 pm, jeffunit wrote: > At 07:23 PM 8/20/2009, NeilBrown wrote: >>On Fri, August 21, 2009 12:01 pm, jeffunit wrote: >> > I had a computer that was running mdadm with no problems. >> > Then the motherboard failed, while the array was in the process of >> > rebuilding. >> > I replaced the motherboard with a new motherboard, >> > and I an trying to rebuild my raid array. >> > >> > I am running mandriva 2009.1 spring (the latest). The kernel is >> 2.6.29.6 >> > I have a 7 drive raid 5 array. >> > I said >> > mdadm --assemble /dev/md0 /dev/sd[bcdefgh] >> > >> > I got the error >> > mdadm cannot open device: /dev/sdb device or resouce busy >> > mdadm device /dev/md0 has no superblock - assembly aborted >> > >> > I then tried to make a degraded array >> > mdadm --assemble /dev/sd[cdefgh] --force >> > That seemed to work, and the array was degraded. >> > >> > I decided to reboot to see if I could get /dev/sdb online. >> > I did, and it didn't help. >> > >> > What can I do to recover my array? >> >> >>You need to find out why sdb is busy. >>Maybe look in /sys/block/sdb/holders. >>Maybe some filesystem on it go automagically mounted by mistake >>or something. >>Maybe send: >> cat /proc/mounts /proc/swaps /proc/mdstat /proc/partitions >> find -H /sys/block/sdb -ls >> >>and we can see if there are any hints. >>Maybe send full kernel logs from the boot sequence as well. > > The 7 drives are dedicated to raid. I checked /etc/fstab, and none of > the raid drives are there. > I have included everything but the kernel logs. > Does any of this help with a path forward? > > thanks, > jeff > > /sys/block/sdb/holders has: > 0 lrwxrwxrwx 1 root root 0 2009-08-20 19:49 md_d125 -> > ../../../../../../../../. > ./../virtual/block/md_d125/ > > (I don't understand this, but I suspect it is the problem, and I > don't know how to fix it) I suspect you are right !! > # cat /proc/mdstat > Personalities : > md_d121 : inactive sdh[6](S) > 732574464 blocks > > md_d122 : inactive sdd[2](S) > 732574464 blocks > > md_d123 : inactive sde[3](S) > 732574464 blocks > > md_d124 : inactive sdc[1](S) > 732574464 blocks > > md_d125 : inactive sdb[0](S) > 732574464 blocks > > md_d126 : inactive sdf[4](S) > 732574464 blocks > > md_d127 : inactive sdg[5](S) > 732574464 blocks > It looks like the boot scripts tried running "mdadm --incremental" on each device and it didn't do the right thing for some reason. This is easy enough to fix in the short term: mdadm -Ss mdadm -A /dev/md0 /dev/sd[bcdefg] You might find that it assembles degraded, and that you need to mdadm /dev/md0 --add /dev/sdb or similar to get it to rebuild. Longer term, you want to make sure it does the right thing at boot time. Maybe you just need to change something in /etc/mdadm.conf. I have no idea how mandriva organised md assembly at boot time so I cannot offer you concrete advice. NeilBrown -- 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
