TFTP Server

[Raymond Bell]

@Tom-Elliott said in TFTP Server:

Typically we see this happening when there’s more than one DHCP server handing out PXE information.

I do have 2 DHCP servers that replicate to each other. But it has always been this way think i should remove pxe 66/67 from one of the DHCP servers and see what happens?

[Tom Elliott]

@Raymond-Bell If one is a redundant to the other, it shouldn’t be a problem I don’t think. You can attempt it though.

[Raymond Bell]

@Tom-Elliott said in TFTP Server:

@Raymond-Bell If one is a redundant to the other, it shouldn’t be a problem I don’t think. You can attempt it though.

@Tom-Elliott Yes it is redundant and like i was saying it has always been this way and never had issue like this one that started this morning.

[Raymond Bell]

@Tom-Elliott That did not work. removing 66/67 from one of the DHCP servers.
So i sat up one dhcp server with 66/67 and made sure that the computer i am booting is pulling from that DHCP and still same issue… I deleted 66/67 off the other dhcp server just to test.
So all it had to pull from was the DHCP server that has the 66/67 on it and it done the same thing asking to enter TFTP server info

[Tom Elliott]

@Raymond-Bell Is this happening on this one machine/model or all machines?

[Raymond Bell]

@Tom-Elliott it is happening on like half of them of course you know users they never turn there computers off lol
But i have tried couple that it was fine on it went right into fog and some that it gives this enter server msg same model and different models

[Tom Elliott]

@Raymond-Bell Is there any possibility of a Rogue DHCP server on your network right now?

[Raymond Bell]

@Tom-Elliott said in TFTP Server:

@Raymond-Bell Is there any possibility of a Rogue DHCP server on your network right now?

Nope Unless one of the new switches is handing out dhcp but i dont see that happening

[Tom Elliott]

@Raymond-Bell Do the new Switches have the proper IP Helpers in place to point at the DHCP server?

I know this is a lot of questions, just thinking all possibilities.

A good way to test is to put a machine that may be on your new switch, and put it in between (before the new switch) and see if all is working properly.

[Raymond Bell]

@Tom-Elliott said in TFTP Server:

@Raymond-Bell Do the new Switches have the proper IP Helpers in place to point at the DHCP server?

I know this is a lot of questions, just thinking all possibilities.

A good way to test is to put a machine that may be on your new switch, and put it in between (before the new switch) and see if all is working properly.

LOL i would but every switch is a new switch and all i done was set up the Vlans on them 40 new switches and new single mode fiber

[Tom Elliott]

@Raymond-Bell Well, if I remember my working in a Central Network area, the only switch that required the IP helper was the “central” switch. So assuming a similar setup (multiple switches, all vlan’s pointing to a Central switch, and that switch communicating with your DHCP server) the IP Helper statement would just need to be setup for each vlan to point to your appropriate dhcp server.

Hopefully this helps.

Also, if you had VLAN specific Boot Options for 66/67, maybe make sure all are setup appropriately. I’m guessing, though, that you have it set as the “global” scope.

[Raymond Bell]

@Tom-Elliott You think i need to turn DHCP replay on on all the switches? or just the core switch?

[Tom Elliott]

@Raymond-Bell Just the “core” switch as I’m guessing all others already know how to route to the core.

[Raymond Bell]

@Tom-Elliott I don’t think it would be the core because the STATE owns the core switch… I may just try just that link port on first switch in line from the Core

[Raymond Bell]

@Tom-Elliott Ok after some testing with 2 computers plugged in to same new switch one works great boots right in to fog the other enter tftp server same computers models same DHCP

[Tom Elliott]

@Raymond-Bell Then I really don’t know what’s going on. I’m sorry man, maybe @Sebastian-Roth or @George1421 can be of some more help. I’m not much on the figuring out side as I used to be, and I’ve given everything I can think of.

Sorry I’m not more useful at this point.

[Raymond Bell]

@Tom-Elliott said in TFTP Server:

@Raymond-Bell Then I really don’t know what’s going on. I’m sorry man, maybe @Sebastian-Roth or @George1421 can be of some more help. I’m not much on the figuring out side as I used to be, and I’ve given everything I can think of.

Sorry I’m not more useful at this point.

Ok thanks man…

i even unplugged the one that was giving me the tftp and plugged it into the port one that is work and same error

[george1421]

@Raymond-Bell TBH I only skimmed this thread so I may have missed something. But my intuition is telling me there is a secondary dhcp server responding to the target computer giving it conflicting pxe boot info.

[george1421]

So lets start with some data collection.

Is the dhcp server, fog server and target computer all on the same subnet/vlan? If so we can use the fog server to help us see what is going on in the pxe booting process.

I have a tutorial here on how to use tcpdump from the FOG server to capture a pcap file of the booting process: https://forums.fogproject.org/topic/9673/when-dhcp-pxe-booting-process-goes-bad-and-you-have-no-clue

If the dhcp server, fog server, and pxe booting client are on the same subnet/vlan then go and grab the pcap and post it here. If any is on a different vlan then give us an idea of your network design so we can come up with a useful plan to collect this information.

[Raymond Bell]

@george1421 yes there are all on same subnet/vlan
Here you go sir0_1542402185175_output.pcap

10.24.32.60 is the ip the host pulled from dhcp