- This topic is empty.
12. May 2009 at 20:08 #2605
I’m trying to push out images I made for my company’s 3150SE and S30 terminals, however I cannot get them to find the WDM server while PXE booting.
The WDM server is on the same core switch as the thin clients, however it is on a different VLAN. Adding IP helper addresses to the thin clients’ VLAN has not worked. The setup gets a little more complicated, however: WDM server’s VLAN uses our AD controller as its DHCP server, whereas the Cisco switch is acting as the DHCP server for the VLAN that the thin clients are on. I tried “hardcoding” the PXE client response into the DHCP pool on the Cisco switch using options 60 and 43 with the WDM server’s IP address in the proper hex format as described in the guide on this website, but it has not worked either.
I did find that when the IP Helper addresses were added or the PXE response and server address were “hardcoded”, the error message I received was “No DHCP or proxyDHCP offers were received”, whereas it would simply display “Exiting PXE ROM” if no helpers or hardcoded response were present.
What can I do to resolve this issue? Any help and/or info would be greatly appreciated.13. May 2009 at 6:27 #15845
That sounds like a real problem.
Can you stop the Cisco DHCP server for a few minutes and check if it would work then? Just to eliminate one of the possible problems.
CG13. May 2009 at 9:09 #15847johnjoreMember
- Total Post: 6
We have the PXE servers on one subnet and clients on 50 other subnets. IP helpers are only used for the DHCP servers, but option 043, 60 and 66 tell the client where the PXE servers are.
In other words PXE and DHCP are on different servers.
With this we now have PXE boot working for V90LEs.
JJ14. May 2009 at 12:23 #15861
I set up a test VLAN with the exact same configuration as my production environment and got the same result when I tried to PXE boot a WinTerm. I then tried removing the Cisco DHCP pool from that VLAN, however it did not change anything other than the WinTerm no longer being able to lease an address after it finished booting from its internal memory.
I’m going to run a packet sniffer and see if I can track the DHCP broadcast from the WinTerm.14. May 2009 at 12:36 #15863johnjoreMember
- Total Post: 6
If it helps, here are my settings:
My option 43 looks like this: (With no spaces) Replace xx with the IP addresses of the PXE boot servers in hex values (Must specify two in my setup)
080B 000E 02 xxxxxxxx xxxxxxxx
0905000e 0252420a 020052ff
Option 60 = “PXEClient”
Option 66 = “PXESERVER.fqdn.net” < your FQDN of main PXE server
(We’re still working on getting two PXE boot servers per subnet to work, one for WDM and one for SMS, but it works fine for one)
Our IP Helpers are only setup to point to the core DHCP servers, which are on different servers to the rest of the infrastructure.
JJ15. May 2009 at 14:29 #15876
Thanks for the responses. I managed to get to the bottom of the issue. It turns out that Spanning Tree Protocol is running on our switches and the ports the WinTerms were connected to were not coming online until after the PXE timed out.
Enabling portfast on those ports solved this problem, and I was finally able to image my S30 and 3150 thin clients.
I still couldn’t get the PXE loader to find the WDM server using IP helpers, but “hard coding” the PXE response and WDM server’s address into the DHCP pool worked perfectly. Luckily, all my thin clients are on their own VLAN, subnet range and DHCP pool, so it won’t affect other PXE clients on my network (and I get the added bonus of the WinTerms not accidentally picking up the Tivoli provisioning manager server anymore.)
- You must be logged in to reply to this topic.