Re: Software RAID function of RH7.2 is not yet reliable.

This is self reply:

I thought the software RAID code were the only reason to get kernel stop.
But, I now think the IDE HDD interface driver could be the reason.

Using the Promise Ultra66 or Ultra100 PCI IDE cards,
connecting the both primary and secondard cable the master's position
the Maxtor 80GB ATA disks, they become /dev/hde and /dev/hdg in RH7.2-IA64.
The shell script like:
---------
#!/bin/sh
 dd if=/dev/hde of=/dev/null &
 dd if=/dev/hdg of=/dev/null &
---------
after it is run little while, it soon drops the kernel. 
The similtaneous access to these HDDs (even without makeing software raid)
is not working correctly. 

It may be that my system has dual CPU is bad for the (Promise's) U-ATA UDMA
interface card driver. Or the IDE interface chip's driver for the Promise 
series card is incomplete.

The simultaneous heavy access to two SCSI disks on the same SCSI bus,
by the experiment, does not seems to drop the kernel.

The simultaneous heavy access to the SCSI and the (Promise interface) 
IDE disk, by the experiment, does not seem to drop the kernel.

The simultaneous heavy access to both (Promise interface) IDE disk, 
on the same card by the experiment, does drop the kernel 100%.

I would go and get other IDE interface card than Promise's 
that can be used on IA-64 system and test on it later.

#
On the boot time, some message like:
creating /etc/fstab, etc, "modprobe block-major-2 failed" or 
some like that message is shown, but dmesg command does not re-display it.
May be I should re-install RH7.2 from scratch again.

H.M.