ipconfig and initrd problems

[archived]

Am I the only one having problems with network booting after the 2.0.40 upgrade?  I've even reinstalled from scratch using the new CDs, and while the install goes ok, I still can't get a remote MD to boot.

I've tried to network boot with both a diskless machine which worked fine with Pluto .39 and a virtual machine under VMWare, which also worked fine.  As far as I can tell, I've set up everything on the main Pluto box ok, as the machines find their kernel and initrd images.  The problem is that the boot scripts stall before mounting the NFS exported directories.

I've attempted to debug by adding "break=modules debug=y" to the kernel command line in pxelinux.cfg, which drops me into a shell so I can figure out what the problem is.  The stopping point seems to be "ipconfig", which should configure the network interface and dump configuration information into /tmp.  All that happens when ipconfig runs is that it continually (every 30s or so) asks for a lease, which it is granted, but it never responds with an ACK.  Eventually it complains about some TX timeout.

I'm going to try hacking the initrd to see if I can at least get the MD to boot and mount the NFS exports, but I can't believe I'm the only one to see the problem.

Alex

[archived]

Ok, so I was the only one.

I've figured out that there's some regression bug in the new kernel which stops the onboard NIC of my EPIA SP 8000 from working, which I've worked around by adding "pci=noacpi" to the kernel parameters.

As for why my VMWare netboot doesn't work, I dunno.  Perhaps some strange firewall issue.