Red Hat AS 4 questions

[Rick Stevens]

Allen, Jack wrote:
> 
> -----Original Message-----
> From: Rick Stevens [mailto:rstevens at vitalstream.com] 
> Sent: Wednesday, August 10, 2005 6:07 PM
> To: Getting started with Red Hat Linux
> Subject: Re: Red Hat AS 4 questions
> 
> 
> Allen, Jack wrote:
> 
>>        I just loaded Red Hat AS 4 on and IBM xServer 366. Everything 
>>when fine. I have 2 disk in Raid 1, that everything was loaded on and 
>>appear as /dev/sdg1 - /dev/sdg9. Then there are 12 other disk, that are 
>>in Raid 1, so the OS sees 6 disk. These 6 disk were all setup with one 
>>partition and appear as /dev/sda1 through /dev/sdf1. I ran pvcreate 
>>/dev/sda1, pvcreate /dev/sdb1 ... Then I ran vgcreate vg01 /dev/sda1 
>>/dev/sdb1. The vgdisplay command shows the correct information, no 
>>problem. Then I ran
>>
>>lvcreate -L 2G -n h0000n0.v00a vg01. This is for and application to 
>>directly access, so I needed to change the owner, group and mode. So I 
>>went to /dev/vg01 and did "ls -l" to see what it currently was. This is 
>>when I found out the name h0000n0.v00a was a symbolic link to 
>>/dev/mapper/vg01-h0000n0.v00a. So my first question is: does this mean I 
>>am running the dev file system or something by default? Or is this the 
>>way that lvm2 works now?
> 
> 
> RH AS4 is a 2.6 kernel, unlike AS3 and 2.1, which were 2.4 kernels.
> The 2.6 kernel uses "udev" to manage device names and such.  What you're
> seeing is how udev does things.  Check the man pages for udev.
> 
> 
>>        Now to my second problem. When I loaded Red Hat AS 3, I don't 
>>remember doing anything special to create the /dev/st* and /dev/nst* 
>>nodes. It seems there was something like 7 of each. Well on the this 
>>system there are none. The man page for st gives an example of "mknod -m 
>>0666 /dev/st0 c 9 0", which I did. No errors. But if I do "mt -f 
>>/dev/st0 status" the error indicates no such device or address. I would 
>>assume this is because the major device number may not be correct. So my 
>>question is: what should it be? Or is it related to the /dev/mapper 
>>question above? As far as the tape drive really being connected to the 
>>system, it shows up during boot when the controller is identified. This 
>>is before Linux boots. I don't remember if it showed up during the Linux 
>>boot. I will have to reboot again to see. I guess another question would 
>>be is all the boot messages from the kernel logged somewhere? I thought 
>>it was, but don't remember where.
> 
> 
> If you check /var/log/boot.log, you should see where the device was
> identified and assigned a name.  You'll probably find it listed as 
> "/dev/tape*" rather than "/dev/st*".  Again, this is udev doing its
> thing and you should familiarize yourself with it.
> 
> The 2.6 kernel is a radical departure from what you're used to.  It's
> a better kernel, it's faster, it's cleaner and it handles memory much
> better.  You'll like it, but there is a bit of a learning curve.
> ----------------------------------------------------------------------
> - Rick Stevens, Senior Systems Engineer     rstevens at vitalstream.com -
> - VitalStream, Inc.                       http://www.vitalstream.com -
> -                                                                    -
> -  You know the old saying--any technology sufficiently advanced is  -
> -               indistinguishable from a Perl script                 -
> -                                 --Programming Perl, 2nd Edition    -
> ----------------------------------------------------------------------
> 
> Thanks for the information Rick. I will have to read up on udev. I just
> rebooted the system before I read this and noticed the first thing that gets
> started is udev. I had found the boot.log file and looked through it. I did
> not see anything about the 2 SCSI controllers the tape drives are connected
> to. There is a tape drive on each controller, one is a DLT1 and the other is
> a 4mm. As I said they show up in the preboot messages, but it looks like the
> kernel does not see them. The controllers are IBM Ultra320 SCSI Controller 2
> by Adaptec. Any Ideas?

Ah.  Sounds like the drivers for the controllers weren't loaded.  Try
doing an "lsmod" and see if any "aic7*" drivers were loaded (there
are two primary Adaptec drivers, "aic7xxx" and "aic79xx").  If not,
try "modprobe aic7xxx" and see if it loads.  If it doesn't, try
"modprobe aic79xx".  One or the other should load.  When they do, check
the "dmesg" command's output and see if the drives show up.  If they do,
then add lines to /etc/modprobe.conf:

	alias scsi_hostadapterX aic7xxx

Replace "X" with the next sequential number from the other
"alias scsi_hostadapter" lines in the file, replace "aic7xxx" with
"aic79xx" if you're using the aic79xx driver, and repeat for each
Adaptec you have.  For example, if you have two RAID controllers in your
machine and two Adaptecs which use aic79xx, you'll have the following
lines in /etc/modprobe.conf:

	alias scsi_hostadapter0 raid-controller-driver-name
	alias scsi_hostadapter1 raid-controller-driver-name
	...
	alias scsi_hostadapter2 aic79xx
	alias scsi_hostadapter3 aic79xx

When you reboot, the drivers _should_ load and "lsmod" should show them.
If they don't load (because you don't immediately start whacking the
tape drives at boot), you may also need to add "modprobe aic79xx" or
"modprobe aic7xxx" to your /etc/rc.d/rc.local file, just to make sure
the drivers get loaded.
----------------------------------------------------------------------
- Rick Stevens, Senior Systems Engineer     rstevens at vitalstream.com -
- VitalStream, Inc.                       http://www.vitalstream.com -
-                                                                    -
-          When all else fails, try reading the instructions.        -
----------------------------------------------------------------------