0

We use WDS to install new workstations, and the newest batch we've gotten are set to use UEFI, so I'm trying to adjust my setup accordingly.

Previously, we had the 66 and 67 options set on our DHCP server for the WDS server and the boot file path, respectively. We did not need option 60, since the WDS server is a different machine than the DHCP server. (Side note: the workstation VLAN is different than the server VLAN.) This worked fine with the "legacy" boot - and still does, if I revert the new equipment. However, I'd like to not have to do those steps on every machine going forward. I tried to put in the UEFI boot path instead, but got an error message 0x102.

As I looked into how to fix that, I'm seeing lots of recommendations to not use the DHCP options, but IP helper addresses. We currently have an IP helper address set on our stack (Cisco Catalysts) to point to the DHCP server. As far as I can see, I can do one of three things:

  1. Replace the existing IP helper for the DHCP server with one for the WDS server. But what will this do to regular (non-PXE boot) DHCP requests?

  2. Leave it, and keep fiddling with the DHCP options. But I haven't seen anything that looks really promising re: a solution, and it's against best practices.

  3. Add a second IP helper address with WDS server (in addition to the existing one for the DHCP server). But from what I can see, each device will send a request to both IP helper addresses, and then basically go with whatever responds first. This doesn't seem like it would produce the desired behaviour.

Sadly, I don't have a parallel test environment I can play on here, plus I'd like to understand the theory behind it, rather than just going off of what I see as I flail. :-) Anyone have any resources they can suggest, and/or thoughts on how to best handle this?

teleute00
  • 347
  • 1
  • 4
  • 16
  • What Windows server version is you DHCP server and what Windows server version is your WDS server? I want to say that there is almost no config changes needed on the DHCP server if you use 2012 R2. Same with the WDS server. Just add the UEFI compatible boot and install images. If you have Windows 10 Pro then you can just rev up a hyper-v VM and test boot with it. As long as it is Gen 2. – Elliot Huffman Jan 17 '18 at 07:06
  • Both are 2012 R2. I did add the UEFI images, but it didn't work. (That's when I got the error message 0x102.) – teleute00 Jan 18 '18 at 07:46
  • Are they retail images? – Elliot Huffman Jan 18 '18 at 07:48
  • No - everything's from volume licensing versions. – teleute00 Jan 18 '18 at 22:28
  • I'm sorry when I'm at retail images I meant untouched straight from Microsoft images. – Elliot Huffman Jan 18 '18 at 22:28
  • Ah - then yes. I just start with the base OS image, and then during deploy use WDS to do app installs and customizations after install. The images remain clean. – teleute00 Jan 18 '18 at 23:42
  • What os are you trying to deploy? Can you also grab the server logs for WDS and a packet capture of the client? Thanks! – Elliot Huffman Jan 18 '18 at 23:50
  • Sorry - had a deadline to get the couple of machines I was working on deployed, so had to work around by changing the boot to Legacy from UEFI. I'll try to come back to this again soon, but might be a few more days. – teleute00 Jan 31 '18 at 00:26

0 Answers0