Raid superblock woes - what keeps trying to move my raid drives at boot?
Hi all,
I have a working raid0 setup, on partitions hdb2, hdc2, and hdd2. However,
every time I reboot, something makes the system look for the raid parts on
the whole drives instead of the partitions.
From dmesg:
> device-mapper: 4.0.5-ioctl (2003-11-18) initialised: dm@uk.sistina.com
> md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
> [events: 00000006]
> md: bind<ide/host0/bus1/target0/lun0/disc,1>
> [events: 00000006]
> md: bind<ide/host0/bus1/target1/lun0/disc,2>
> [events: 00000006]
> md: bind<ide/host0/bus0/target1/lun0/disc,3>
> md: ide/host0/bus0/target1/lun0/disc's event counter: 00000006
> md: ide/host0/bus1/target1/lun0/disc's event counter: 00000006
> md: ide/host0/bus1/target0/lun0/disc's event counter: 00000006
> md: device name has changed from ide/host0/bus0/target1/lun0/part2 to ide/host0/bus0/target1/lun0/disc since last import!
> md: device name has changed from ide/host0/bus1/target1/lun0/part2 to ide/host0/bus1/target1/lun0/disc since last import!
> md: device name has changed from ide/host0/bus1/target0/lun0/part2 to ide/host0/bus1/target0/lun0/disc since last import!
I have to do the following after every boot:
mdadm --stop /dev/md0
mdadm -Cv /dev/md0 -l0 -n3 -c32 /dev/hdb2 /dev/hdc2 /dev/hdd2
then restart my vg and mount my partitions.
After recreating the array, both mdadm -E /dev/hdb and mdadm -E /dev/hdb2
give the same result (I think they look at the same superblock), ending in:
> 0 0 3 66 0 active sync /dev/hdb2
> 1 1 22 2 1 active sync /dev/hdc2
> 2 2 22 66 2 active sync /dev/hdd2
Anyone have an idea why the system wants to use the whole drives at
the next boot??
Thanks,
Peter
Reply to: