The issue may be related to the kernel module include in the client rpm's (nbd.ko) if its not found in on the client side it won't be able to net boot. First mount up your image file as a loopback, chroot into it then check to see if nbd.ko exists in your modules directory for your kernel:
# ls /lib/modules/`uname -r`/kernel/drivers/block/nbd.ko
Also make sure you have the nbd rpm installed on the server
Note that nbd-3.3 hasn't been tested yet, nbd-2.9.20 is the current supported version.
From: Jaap Bril <jaap jbril net>
Reply-To: "Support schools." <k12osn redhat com>
Date: Fri, 02 Aug 2013 19:04:56 +0200
To: "Support schools." <k12osn redhat com>
Subject: [K12OSN] nbd-server on Centos 6.4
I'm having a problem getting ltsp to work with nbd-server on Centos 6.4.
dhcp (different server) is ok; pxelinux.0 as well as vmlinuz.ltsp and initrd.ltsp arrive on the client.
nbd-server starts and seems ok (expected client-ipaddress, image-file /opt/ltsp/images/i386.img and file-size mentioned by xinetd)
After that things get hectic (I took quiet off) and eventually go wrong;
The client-screen blanks and nbd-server stops (exit=1).
I think that the image is ok; I ran ltsp-build-client yesterday and ltsp-update-image today.
How can I definitively check that the image is ok?
Can I pause the client to get time to read the text?
Any other suggestions?