[NCLUG] Adaptec Raid woes, cont..

[jbass at dmsd.com]

	Thanks for the help earlier.  I ran the followning commands, and all 
	completed without error: 

	# fdisk /dev/sda1
	  IN fdisk, I made sure there were no pre-existing partions left (deleted 
	them), than selected to create one primary partition, partition 1.  I wrote 
	this to the drive. 

Creating a partition table inside partition 1 of /dev/sda is confusing at
best. Doing so directly on the drive, aka /dev/sda, is a bit more meaningful.

	# mkfs -t ext3 /dev/sda1 

	This completed without error.  The activity lights on the drive array 
	indicated that writes were occuring on all disks. 

	I then ran the following, and received these errors.  Any ideas? 

	[root at Leah mnt]# mount -t ext3 /dev/sda1 /mnt/raid
	EXT3-fs error (device sd(8,1)): ext3_check_descriptors: Block bitmap for 
	group 34 not in group (block 0)!
	EXT3-fs: group descriptors corrupted !
	mount: wrong fs type, bad option, bad superblock on /dev/sda1,
	      or too many mounted file systems
	[root at Leah mnt]# 

	Thanks - Erich Hennig

something is corrupting the filesystem ... which is obvious. If you are consistant
about using the whole drive, or with partitions, and the errors still occur, then
it's obvious there is a problem with one or more drives, the raid controller, it's
driver, or the system you are using it in - and in some cases, maybe the cables or
termination.

start with zeroing the drive array as described (dd if=/dev/zero of=/dev/sda bs=1024k)
then mkfs the drive without partitions (mkfs -t ext3 /dev/sda) and try to mount it.
If you get errors doing that, then it's highly probable that your array subsystem is
unstable .... start by testing the drives outside the raid controller, then test the
drives as an array.

John