Problems with booting WinPE2 via linux pxe

[Parak]

Hi all,

I'm having a bizarre problem with getting the aformentioned combination to work...again. I say again because it in fact worked just fine for a few days, using pxeboot.n12 and bootmgr.exe in the tftpboot, and the rest of the required files in tftpboot/Boot. Then, it just for no reason (of course there has to be one, but I'm completely stumped) stopped working, and only the Winpe booting - I use this server for kickstarts which continued to work just fine. Digging through tcpdump shows that everything is loading fine in the following order:

pxelinux.0 -> menu.c32 -> pxeboot.n12 -> bootmgr.exe -> \Boot\Fonts\wgl4_boot.ttf -> boot.ini (doesn't exist) -> request for pxe (not applicable).

At this point the client (.101) hits a brick wall when it requests \Boot\BCD from the tftp server (.2) (tcpdump -vv -i any):

IP (tos 0x0, ttl  20, id 512, offset 0, flags [none], proto: UDP (17), length: 54) 192.168.1.101.9130 > 192.168.1.2.tftp: [udp sum ok]  26 RRQ "\Boot\BCD" octet tsize 0
arp who-has 192.168.1.101 tell 192.168.1.2
arp who-has 192.168.1.101 tell 192.168.1.2
IP (tos 0xc0, ttl  64, id 53633, offset 0, flags [none], proto: ICMP (1), length: 71) 192.168.1.2 > 192.168.1.2: ICMP host 192.168.1.101 unreachable, length 51
IP (tos 0x0, ttl  64, id 33934, offset 0, flags [none], proto: UDP (17), length: 43) 192.168.1.2.42686 > 192.168.1.101.9130: [udp sum ok] UDP, length 15
IP (tos 0xc0, ttl  64, id 53634, offset 0, flags [none], proto: ICMP (1), length: 71) 192.168.1.2 > 192.168.1.2: ICMP host 192.168.1.101 unreachable, length 51
IP (tos 0x0, ttl  64, id 33935, offset 0, flags [none], proto: UDP (17), length: 43) 192.168.1.2.42686 > 192.168.1.101.9130: [udp sum ok] UDP, length 15
IP (tos 0xc0, ttl  64, id 53635, offset 0, flags [none], proto: ICMP (1), length: 71) 192.168.1.2 > 192.168.1.2: ICMP host 192.168.1.101 unreachable, length 51
IP (tos 0x0, ttl  64, id 33936, offset 0, flags [none], proto: UDP (17), length: 43) 192.168.1.2.43050 > 192.168.1.101.9130: [udp sum ok] UDP, length 15
IP (tos 0x0, ttl  20, id 513, offset 0, flags [none], proto: UDP (17), length: 54) 192.168.1.101.9130 > 192.168.1.2.tftp: [udp sum ok]  26 RRQ "\Boot\BCD" octet tsize 0
arp who-has 192.168.1.101 tell 192.168.1.2
arp who-has 192.168.1.101 tell 192.168.1.2

<and so on until it tries 8 times to get it>

The arp and ICMP unreachables are odd, as they only occur here and not during other parts of the boot process - for example tftp RRQ for bootmgr.exe didn't have them. If I tcpdump only on 192.168.1.2 (instead of any), I only see the arp requests and the RRQs for \Boot\BCD again with no answer of any sort from the server. I can't ping the client during any step of the PXE process (should I be?). TFTPD is not showing any errors, and logs the appropriate request for RRQ \Boot\BCD along with remapping. I can also use a tftp client from a third box and request that file in octet mode with no problems.

The end result is that after 8 tries to get \Boot\BCD, bootmgr shows 0xc0000001 "An error occurred while attempting to read the boot configuration data". I've tried a number of different tftp servers, replaced the client completely, and messed with a number of dhcpd/tftpd settings, all to no avail. Hopefully someone that runs a similiar configuration can shed some insight or a critical look as to what I'm missing

Thanks!