 | | | Red Hat diskless feature | Red Hat diskless feature 2005-03-15 - By JJ
Back
With the DHCP config the same, I rsynced the /root directory from my WBEL3
server to a FC3 box, no matter what my config was I could not get it to
boot, I always got a init error, it looked like it ran out of ramdisk by
1000 times..., and a try passing init=...
I copied the /tftpboot/linux-install/pxelinux.0 from my functioning WBEL3
box to the FC3 box and my diskless booted the first time...
I'll go look for more info from the FC3 side but, that fixed the problem
from trying to boot from an FC3 box.
J.J.
-- --Original Message-- --
From: taroon-list-bounces@(protected) [mailto:taroon-list-bounces@(protected)]
On Behalf Of JJ
Sent: Monday, March 14, 2005 8:16 AM
To: 'Discussion of Red Hat Enterprise Linux 3 (Taroon)'
Subject: RE: Red Hat diskless feature
Holy cow... I just got it working...
My setup...
-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
DHCP Server, TFTP Server: WBEL3 latest patches
Initial Image Workstation: WBEL3 latest patches
Diskless Client: VMWare Guest
-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
/etc/dhcpd.conf (changes)
allow booting;
allow bootp;
host wbel3-nfs-1 {
hardware ethernet 00:0c:29:cc:54:f2;
fixed-address 204.72.141.55;
next-server 204.72.141.1;
if substring(option vendor-class-identifier, 0, 9) = "PXEClient" {
filename "linux-install/pxelinux.0";
}
}
-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Generated pxeboot file from redhat-config-netboot:
/tftpboot/linux-install/pxelinux.cfg/CC488D37
default WBEL3
label WBEL3
kernel WBEL3/vmlinuz
append console=tty0 initrd=WBEL3/initrd.img root=/dev/ram0
init=disklessrc NFSROOT=204.72.141.1:/opt/wbel3-nfs-boot ramdisk_size=10000
ETHERNET=eth0 SNAPSHOT=wbel3-nfs-1
-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
/etc/exports
/opt/wbel3-nfs-boot/root *(ro,no_root_squash,sync)
/opt/wbel3-nfs-boot/snapshot *(rw,no_root_squash,sync)
-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
I booted the VM, tail -f /var/lib/dhcp/* to make sure the VM got the
appropriate ip address.
Mine initially came up with an error can't find grep in the disklessrc ( I
did not have busybox & buysbox anaconda on the box I copied the data off
of... ( rsync -a -e ssh host:/ /opt/wbel3-nfs-boot/root ) )
I chroot'd the root, installed the busyboxes, then used a cool utility,
http://sial.org/howto/linux/initrd/ to unpack the initrd, I copied the
busybox into /sbin and re packed the initrd, moved the new image to the
/tftpboot/linux-install/WBEL3 directory...
This time the box came up with no problems... I can't believe it...
I will replicate this with RHEL3 at work today if I get a chance, but for
sure by the end of the week...
My reason for trying this, is I have an IBM BladeCenter at work ( they gave
it to us free with the M.F. purchase) without disk... My network guys gave
me one pair of fiber from the lab down stairs ( only place for 220, damn ibm
) to the upstairs lab where the MSA1500 sits... I have one of the quad
blades remote booting off the SAN. But I have no disk for the other
blades...
Untill now....
This may not help a bunch, but I just wanted to let you know it's
possible...
J.J.
-- --Original Message-- --
From: taroon-list-bounces@(protected) [mailto:taroon-list-bounces@(protected)]
On Behalf Of NICOLAS GOBET
Sent: Friday, March 11, 2005 8:31 AM
To: taroon-list@(protected)
Subject: Red Hat diskless feature
Hello,
I have 8 Dell PowerEdge 1850 to install. It's a good occasion to try the
diskless feature of redhat.
I followed this documentation:
http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/sysadmin-guide/c
h-diskless.html
But there is a problem:
when the pxe client boots, it makes a kernel panic, because of no init
found.
The server is a Dell Poweredge with Red Hat ES 3 (192.168.0.20)(up to
date) and the nodes will have a RedHat WS 3 distribution.
The setup of the first node works great and I make an image using:
rsync -v -a -e ssh 192.168.0.198:/ /diskless/RHEL3-WS/root/
--exclude="/proc/*" --exclude="/sys/*"
Then I have made the nfs exports:
/diskless/RHEL3-WS/root/ *(ro,sync,no_root_squash)
/diskless/RHEL3-WS/snapshot/ *(rw,sync,no_root_squash)
I can acces these nfs shares with each computer on my lan.
The the dhcpd sends what's needed, exactly as described in the doc.
The PXE client boots, gets the pxelinux.0 (given by the setup of Red Hat)
Then it looks in /tftpboot/pxelinux.cfg, trying to find a file with its MAC
Address and finally loads the client file corresponding to it's ip:
#beginning
default rhel-3-ws
label rhel-3-ws
kernel rhel-3-ws/vmlinuz
append initrd=rhel-3-ws/initrd.img root=/dev/ram0 init=disklessrc
NFSROOT=192.168.0.20:/diskless/RHEL3-WS ramdisk_size=10000 ETHERNET=eth0
SNAPSHOT=192.168.0.220 NFSROOT=192.168.0.20:/diskless/RHEL3-WS
init=disklessrc root=/dev/ram0
#end
As written in this file, the client looks in ...../rhel-3-ws and it loads
the vmlinuz and the initrd.img generated by the Red hat tool.
This works correctly but right after the load of the nfs shares appears the
kernel panic:
here are the interesting lines:
RAMDISK: Compressed image found at block 0
VFS: Mounted root (ext2 filesystem) readonly.
Freeing ununsed kernel memory: 224k freed Kernel panic: No init found. Try
passing init= option to kernel.
It sounds like the nfs share is mounted, but the kernel doesn't see anything
on the "partition".
I was able to mount the nfs share with the cd installation.
Has somebody an opinion about this? Where did I make mistakes?
Thanks
Nicolas
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
Earn $52 per hosting referral at Lunarpages.
|
|
|