System: IBM eServer, dual 1.266, 2GB RAM, HP SmartARRAY 641 w/ 6 36GB in RAID5
VL Version: 6 Standard Gold
After about an hour of messing with the halts during GUI setup and the failure at LILO setup, I've resorted to posting.
The install process runs fine until after partitioning. As a side note, GParted took at least 180 sec. to format a 20GB volume with ext3. When the system attempts to detect the partition scheme, it halts when the scripts that put the $size and $sw variables run. With the CCISS driver, it places the drives (and partitions) in the /dev/cciss/ directory like this:
c
xd
y for the disk (where 'x' is the controller number and 'y' is the logical disk number)
c
xd
yp
z for the partitions (where 'z' is, you guessed it, the partition number)
I assume the scripts to set $size and $sw expect partitions to be like /dev/hda1 and /dev/sda3, and so don't like the extra characters, because a "# ps -ef" shows me
sh -c export size=$(parted /dev/cci print | grep ^'ss/c0d0p1 '); echo $size | cut -f 4 -d ' '
and I assume it should be
sh -c export size=$(parted /dev/cciss/c0d0p1 print | grep ^' '); echo $size | cut -f 4 -d ' '
If I kill those scripts (one for each partition), the system proceeds through the install. Then we get to the LILO install. After killing the same scripts, LILO errors out (and forgive me for not having the EXACT error message) about a variable being invalid, most likely to do, again, with the /dev/cciss nomenclature.
Can anybody give me an idea as to where those scripts are so I can modify them before they execute? The "linux" option at boot time can't even find the drives, so I'm a bit SOL there, as well...
--
Andrew