Hi,
I've got three RAID volumes spread across the disks of a
sparcstorage array 100x on an ultra2 running RH6.1/sparc
(2.2.12 kernel / raidtools 0.90).
When I first signed onto this mailing list, I had a problem with
the disklabels for my partitions being eaten away at reboots. I
was told at the time to start all partitions at cylinder 1 or
above to prevent this. This appeared to work for the immediate
going.
However, I'm still seeing this problem has resurfaced, and half
the drives in the volumes have lost their disklabels/partition info
again. I'm wondering whether there's still something else I needed
to do?
Here's the typical disk layout I've been using:
root@lnxserv tech]# /sbin/fdisk -l /dev/sdz
Disk /dev/sdz (Sun disk label): 19 heads, 80 sectors, 2733 cylinders
Units = cylinders of 1520 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sdz3 u 0 2733 2077080 5 Whole disk
/dev/sdz5 1 2733 2076320 fd Unknown
[root@lnxserv tech]#
Does the "start at cylinder 1" rule apply to the Sun 'whole disk'
entry too?
I'd appreciate any pointers anyone may have.
Thanks,
-Robb
