BLAG

BLAG Forums
It is currently Sun Dec 21, 2014 7:32 pm

All times are UTC




Post new topic Reply to topic  [ 10 posts ] 
Author Message
PostPosted: Wed Mar 19, 2008 3:59 pm 
Offline

Joined: Thu Jul 13, 2006 5:01 pm
Posts: 509
Location: Boston, MA, USA
This is probably out-of-scope for this forum, but as I am running BLAG 70001 on my NFS/Kickstart server, I'll post here anyway. Heck, HP and RedHat have been useless to me, and BLAG quite the opposite, so there!

I'm using HP's SmartStart Toolkit to create an automated Red Hat install process, but can't seem to initiate a kickstart build due to an apparent failure to load network drivers upon boot. SmartStart successfully loads a saved BIOS and hardware configuration and then creates an initial MBR and boot partition on the hard drive of an HP Proliant DL585. The boot loader (via syslinux), initrd.img, vmlinuz and syslinux.cfg are all loaded to this new partition during a pre-boot phase, and everything seems to start out normally upon reboot. The QLogic, USB and cciss drivers are all loaded, but instead of invoking kickstart, I am prompted for keyboard language, installation method and network configuration info. All of this is in the kickstart config file, which evidently is not being seen due to a failure to connect to the network.

My initial syslinux.cfg file looks like this:

Code:
say Performing RedHat Kickstart Installation
default ks
prompt 1
timeout 150

label linux
  kernel vmlinuz
  append initrd=initrd.img lang=en devfs=nomount ramdisk_size=9216

label ks
  kernel vmlinuz
  append initrd=initrd.img lang=en devfs=nomount ramdisk_size=9216 init=/sbin/init root=/dev/nfs ip=:192.168.5.10:192.168.5.252:255.255.0.0:::'bootp' ks=nfs:192.168.1.43:/home/pablo/tmp/ks/ha2_net.ks.cfg ksdevice=eth2 network=1


As there is very little documentation available on this sort of thing, I've basically slapped together the second "append" line from a number of sources. In a perfect world this would be merely a syntax error, but I don't think I'm even getting that far. In any event, an extra set of eyes would be greatly appreciated. Thanks!


Top
 Profile  
 
 Post subject:
PostPosted: Wed Mar 19, 2008 4:11 pm 
Offline
Site Admin

Joined: Sun Mar 14, 2004 3:17 pm
Posts: 4492
Location: Loveland, Colorado, USA
Can you ping the box? That will simply tell you if net comes up or not.

I do network installs, but generally via ftp/http using a DHCP and tftp server.

dhcpd.conf snippet (putting in your own IPs and MAC addr, where the .5 box is tftp/dhcp server):
Code:
ddns-update-style none;
allow booting;
allow bootp;

...

       host blagbox {
               fixed-address 192.168.1.3;
               hardware ethernet 00:00:00:00:00:00;
               filename "pxelinux.0";
               next-server 192.168.1.5;
}


/tftpboot/pxelinux.cfg/default:
Code:
default linux text
prompt 120
label linux text
  kernel f8/vmlinuz
  append load initrd=f8/initrd.img


So you could substitute in the place of "linux text" your kickstart info and such.

I know it's not exactly what you were looking for, but I thought i'd offer it up anyway...

Good luck,

-Jeff


Top
 Profile  
 
 Post subject:
PostPosted: Wed Mar 19, 2008 5:31 pm 
Offline

Joined: Thu Jul 13, 2006 5:01 pm
Posts: 509
Location: Boston, MA, USA
jebba wrote:
Can you ping the box? That will simply tell you if net comes up or not.


I can't ping the box or ping from it when I shell out of anaconda. The error I see on the console is "Network is unreachable". Anaconda loads a number of drivers, but I never see the comforting "Trying to start Network" screen. Instead I'm dropped into a prompt and out of the automation loop. :^(

jebba wrote:
I do network installs, but generally via ftp/http using a DHCP and tftp server.

dhcpd.conf snippet (putting in your own IPs and MAC addr, where the .5 box is tftp/dhcp server):
Code:
ddns-update-style none;
allow booting;
allow bootp;

...

       host blagbox {
               fixed-address 192.168.1.3;
               hardware ethernet 00:00:00:00:00:00;
               filename "pxelinux.0";
               next-server 192.168.1.5;
}


/tftpboot/pxelinux.cfg/default:
Code:
default linux text
prompt 120
label linux text
  kernel f8/vmlinuz
  append load initrd=f8/initrd.img


One problem with using dhcpd and tftp is that I'd be setting up a dhcpd server that conflicts with the corporate Windoze box. I can get around this by setting it up on my BLAG laptop and going through a network switch - essentially creating a private "LAN for two" - but I had hoped I could get this hard disk-based approach to work.

jebba wrote:
So you could substitute in the place of "linux text" your kickstart info and such.

I know it's not exactly what you were looking for, but I thought i'd offer it up anyway...

Good luck,

-Jeff


The reply is most welcome. Thanks!

Paul


Top
 Profile  
 
 Post subject: update
PostPosted: Thu Mar 20, 2008 5:19 pm 
Offline

Joined: Thu Jul 13, 2006 5:01 pm
Posts: 509
Location: Boston, MA, USA
I've tried implementing a network boot using a Linksys hub to connect my BLAG laptop (which is functioning as the DHCP/Kickstart server) to an HP Proliant DL585 server. The DHCP client manages to find the server and attempts to connect via TFTP. At this point the process stalls with the following error message:

Code:
PXE-E11: ARP timeout


So far all I've been able to determine is that I might need to set "Option 60" to account for the fact that my DHCP server and bootp server are on the same box, but I haven't found the correct syntax to do this. Stay tuned...


Top
 Profile  
 
 Post subject:
PostPosted: Fri Mar 21, 2008 2:02 pm 
Offline
Site Admin

Joined: Sun Mar 14, 2004 3:17 pm
Posts: 4492
Location: Loveland, Colorado, USA
At one point I had to add this to my dhcpd/tftp server to get it to work in some setups (i think it was with openbsd IIRC). Perhaps you could explore adding these options:

Code:
ddns-update-style none;
allow booting;
allow bootp;


and

Code:
server-identifier 192.168.1.5;
next-server 192.168.1.5;


-Jeff


Top
 Profile  
 
 Post subject: another update
PostPosted: Fri Mar 21, 2008 5:01 pm 
Offline

Joined: Thu Jul 13, 2006 5:01 pm
Posts: 509
Location: Boston, MA, USA
I was able to get past the "PXE-E11: ARP timeout" error by swapping out the LinkSys hub in favor of a Netgear switch. This enabled the PXE client to get an IP address from the DHCP server, after which it duly executed /tftpboot/pxelinux.cfg/default and started booting. It choked trying to invoke kickstart, however. After loading all of the modules - including network drivers tg3 and e1000 - I was once again dropped into anaconda and prompted for network information. The console output follows:

Code:
* load module set done
* getting kickstart file
* sending dhcp request through device eth2
* waiting for link...
* 0 seconds.
* running dhcp for eth2
* pump told us: No DHCP reply received
* eth2 isn't a wireless adapter


This seems bizarre, since DHCP did get me to the /tftpboot directory on the DHCP server, and did so through eth2 - which is indeed not a wireless adapter.

My dhcpd.conf looks like this:

Code:
allow booting;
allow bootp;
ddns-update-style none;
deny unknown-clients;
authoritative;

subnet 192.168.37.0 netmask 255.255.255.0 {

group {
        filename "pxelinux.0";
        server-identifier 192.168.37.2;
        next-server 192.168.37.2;

        host clone {
                hardware ethernet 00:10:18:10:67:CA;
                fixed-address 192.168.37.3;
        }
}
}


and /tftpboot/pxelinux.cfg/default:

Code:
say PXE kickstart configuration file
default ks
prompt 1
timeout 300
label linux
  kernel vmlinuz
  append initrd=initrd.img lang= devfs=nomount ramdisk_size=9216
label ks
  kernel vmlinuz
  append ksdevice=eth2 ks=nfs:192.168.37.3:/tftpboot/ha2_net.ks.cfg initrd=initrd.img network=1


Top
 Profile  
 
 Post subject:
PostPosted: Fri Mar 21, 2008 7:24 pm 
Offline
Site Admin

Joined: Sun Mar 14, 2004 3:17 pm
Posts: 4492
Location: Loveland, Colorado, USA
Try leaving a running ping to see if/when it brings up the address you want.

And in the 'default' file, try eth1, eth0, etc as it may be choosing a different order for ethN at that point than it did when it first grabbed the address.


Top
 Profile  
 
 Post subject:
PostPosted: Mon Mar 24, 2008 3:04 pm 
Offline

Joined: Thu Jul 13, 2006 5:01 pm
Posts: 509
Location: Boston, MA, USA
You are exactly right - except that this DL585 has 5 NICs (2 on the mother board). I tried all 5, and wouldn't you know that eth4 seemed to be the magic port...


Top
 Profile  
 
PostPosted: Wed Mar 11, 2009 10:48 am 
Offline

Joined: Wed Mar 11, 2009 5:42 am
Posts: 2
in the subject a lot of debris. sorry


Top
 Profile  
 
 Post subject: interesting topic
PostPosted: Fri Mar 13, 2009 8:51 am 
Offline

Joined: Thu Mar 12, 2009 9:34 am
Posts: 2
:))))


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 10 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 0 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group