TFTP only working with certain VLANs.
-
@zpoling said in TFTP only working with certain VLANs.:
“Please enter tftp server:”
This is telling me that your dhcp server is not sending out the dhcp option 66 correctly -OR- you have 2 dhcp servers responding to the target computer (like in a primary and secondary configuration) where the second dhcp server doesn’t have the options set correctly.
FWIW: This issue is a networking infrastructure one and not specifically related to FOG.
-
@george1421 I don’t know how it couldn’t be sending the option out properly. If I have them set globally and every subnet works besides one of them, I don’t know what it could be. Unless someone happened to plug in a DHCP server enabled device to that subnet, but there doesn’t appear to be any evidence of that anywhere.
I’m sure this isn’t fog, but I’m still at a loss.
-
@zpoling First understand I’m only suggesting based on past experience. The system prompting for a tftp server in your screen shot is at an interesting point.
Initially the boot server needs to have been originally detected correctly because at the point its asking for the tftp server address that is INSIDE iPXE. So it worked the first time and now the second time the dhcp address is discovered its not detecting option 66 correctly. That (based on past experience) tells me there are more than 1 dhcp server responding.
Using wireshark to grab the pcap connected to the vlan that is in error, then pxe boot a target computer. That will tell us what is flying down the wire. Since its only one vlan I would focus on that vlan because there is something going on that is not expected.
-
@george1421 I would love to send you a pcap of the issue, but it’s no longer happening. I’ve tested several work stations a couple dozen times. They’ve all been fine.
So the VLAN 3 pcap was not enough to see what was going wrong?
-
@zpoling said in TFTP only working with certain VLANs.:
So the VLAN 3 pcap was not enough to see what was going wrong?
Was the pcap taken from a computer connected to vlan 3 while it was doing the asking for tftp server question? It looks like it was taken from the fog server.
-
@george1421 No, it was a TCPdump from the fog server.
-
@zpoling That is what I thought.
Without going to deep into this. DHCP uses broadcast messages to communicate. If there is a router between the system doing the pcap collection and the pxe booting computer we will not see the dhcp communications. This is the bits that are missing, what the client computer was being told.
With a pcap from vlan 3 we should have seen the complete dhcp sequence of discover, offer, request, ack.
-
@george1421 Here you go. Taken from my computer on VLAN 3, attempting a pxe boot on the computer that was having an issue. Though like I said, it seems to be working fine now.
0_1541174066153_vlan3computer.pcap
The computer in question is 10.32.10.118, DHCP server is 10.32.0.224, and the fog server is 10.32.0.31.
-
@george1421 Don’t spend too much time on it if at all since it’s working now. I’d love to know why it wasn’t working, but I assume those packet captures aren’t going to show jack since they weren’t captured during the issue’s occurrence.
-
@zpoling DHCP handshake looks good in the latest PCAP but as you said we won’t find out what was going on when it failed. More often than not those kinds of issues keep coming back until you really find and fix them. Make sure to capture the traffic when you see it happening next time.
Marking this as solved for now.
