My Redhat 6.1 system contained two IDE disks, hda and hdc, set up as a RAID-0 system, and a single SCSI disk with the OS on it. While attempting an update to RedHat 6.2, the installation procedure attempted to mount both /dev/hda5 and /dev/hdc5 as normal ext2 file systems. This failed, but in the process the RAID system was marked as unclean, requiring a lengthy fsck on first boot. Fortunately no data were lost, but I doubt this can be guaranteed. I'd think the tool which attempts to mount a file system should be changed so as to not to touch the disk if there is not a usable ext2 file system on it.
What are the partition types for these partitions - they should be type 0xfd. Also - are these partitions listed in your /etc/fstab? They should not be if they are part of a RAID array.
May be related to bug number 16832.
(hanging head in shame) No, the partitions aren't type 0xfd, they're type 83. I was _sure_ I'd set them to 0xfd when I created the RAID array, since I carefully followed the instructions in the HOW-TO, but obviously I forgot a crucial "w" to fdisk at some point. /dev/hda and /dev/hdc are not listed in /etc/fstab, though. Will any harm be done if I unmount the filesystem, change the partition type on both disks to 0xfd, and remount it?
I would backup your important data before changing the partition types. I think it should work but don't hold me to that.
*** Bug 53859 has been marked as a duplicate of this bug. ***
*** Bug 54754 has been marked as a duplicate of this bug. ***