[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

2.2.19/raidtools2 not cooperating



I'm setting up a potato system to boot off a software RAID, using a
2.2.19 kernel and raidtools2.  The kernel was built with

CONFIG_BLK_DEV_MD=y
CONFIG_MD_MIRRORING=y
CONFIG_MD_RAID5=y

which should be the appropriate settings.  It's currently installed
on a single disk and I'm trying to use the trick mentioned in the
software-RAID-howto of listing the active disk as a failed drive when
creating the array, then copying my data onto the (degraded) md
device, then using raidhotadd to bring the array up to full
redundancy.

I've done this before without incident, but today, it just doesn't
seem to like me...

--- cut here ---
# cat /etc/raidtab 
raiddev /dev/md0
        raid-level              1
        nr-raid-disks           2
        nr-spare-disks          0
        persistent-superblock   1
        chunk-size              4

        device                  /dev/hdg1
        raid-disk               0
        device                  /dev/hde1
        failed-disk             1

# mkraid /dev/md0
handling MD device /dev/md0
analyzing super-block
disk 0: /dev/hdg1, 1999840kB, raid superblock at 1999744kB
disk 1: /dev/hde1, failed
mkraid: aborted, see the syslog and /proc/mdstat for potential clues.
# cat /proc/mdstat 
Personalities : [3 raid1] [4 raid5]
read_ahead not set
md0 : inactive
md1 : inactive
md2 : inactive
md3 : inactive
--- cut here ---

Nothing is sent to syslog when mkraid is run, so no "potential clues"
exist.

Another interesting note is that, although hde and hdg claim to be
identical drives in their labeling, cfdisk says that hde reports
itself as having 255 heads/4865 cylinders and hdg reports 16
heads/77536 cylinders.  I've made sure that all partitions on hdg are
slightly smaller than their corresponding partitions on hde, though,
so the difference in reported geometry doesn't seem like it should
cause any problems.  The partitions on hdg are of type 0xfd ("Linux
raid autodetect") and those on hde are 0x83 ("Linux ext2") for now,
but I intend to change them to 0xfd before adding them to the array.

That pretty well covers everything...  Have I missed any details or
made any obvious mistakes?

-- 
When we reduce our own liberties to stop terrorism, the terrorists
have already won. - reverius

Innocence is no protection when governments go bad. - Tom Swiss



Reply to: