SCIENTIFIC-LINUX-USERS Archives

August 2008

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Tom Rockwell <[log in to unmask]>
Reply To:
Tom Rockwell <[log in to unmask]>
Date:
Mon, 4 Aug 2008 23:22:14 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (77 lines)
Aldo Saavedra wrote:
> Hi,
>
> I was wondering if someone would an insightful comment.
>
>  From reading past posts I managed to get a node to boot from a server
> using PXE+DHCP+HTTP
>
>
> The image resides in /tftpboot and the configuration is
> default SL45
> serial 0,9600n8
> timeout 100
> prompt 1
>
> label localboot
>   LOCALBOOT 0
>
> label SL45
>   kernel SL45-vmlinuz-boot
>   append console=ttyS0,38400 console=tty0 load_ramdisk=1
> initrd=SL45-initrd.img-boot root=/dev/sda1
>
>
>  the node finds it and installs SL 45
> and reboots...  two questions
> 1) If this was a diskless node.. rebooting would not be helpful so how
> is that installation
> done. (I suppose I would need to remove the root=/dev/sda1 but does SL
> installation always
> needs to be reboot?)
>   
The anaconda installer doesn't work for diskless nodes.  For a diskless 
node, you need to have the system stored somewhere.  Usually it is 
accessed via NFS.  The process is basically to take an already installed 
system and copy the files to the NFS accessed area.  Then you PXE boot 
the host and tell it to mount / from the NFS area.  There are various 
how-to's about this process.
> 2) If it installs to the hard disk.. then how can I change the default
> PXE configuration so it performs a localboot the next time. In this case
> while it was powering down I changed the file so the default would be
> localboot .
>   
You have to get the pxelinux config file changed somehow (or change the 
node BIOS to boot the harddrive before PXE).  Cluster software packages 
such as ROCKS or xCAT provide a mechanism to do this automatically.  (A 
server runs on front-end waiting for indication from client that the 
install has finished, and then the change is made.)

> This did the trick, and the machine started booting from the disk but
> then an error CRC error made the system halt while it was uncompressing
> the linux image (the correct one 2.6.9.55) .
> Any ideas?
>   

Might be a hardware issue, has this node been reliable?  On reboot, do 
you get the exact same error?

Good luck,
Tom Rockwell
Michigan State U.
> Cheers,
> Aldo
>
>
> --
> --
> Aldo F. Saavedra
> Phone: +61 (0) 2 9351 5970
> Fax  : +61 (0) 2 9351 7726
> Email: [log in to unmask]
> Mail : High Energy Physics
>        School of Physics A28
>        University of Sydney NSW 2006
>        Sydney, Australia
>   

ATOM RSS1 RSS2