Please Enter TFTP Server?

[monasmith529]

Server

• FOG Version: 1.3.5 RC-12
• OS: 16.04.2

Client

• Service Version: 0.11.10
• OS: Windows 10

Description

When attempting to PXE, I am getting a prompt to enter tftp server. I have tried entering my fog server IP, but it fails.

[Junkhacker]

what are you using as your dhcp server?

[monasmith529]

@Junkhacker Windows DHCP.

[george1421]

I have see this when the dhcp server is not supplying the dhcp option 66 {next-server} option to the client, the iPXE kernel will prompt for this value. Just be aware this is not a FOG issue at this point. Your target system hasn’t made it to the FOG server to do anything just yet.

I have also see this where you have another imaging function running on the same network (like WDS) and the client is getting conflicting information from dhcp.

[george1421]

@george1421 Wait now that I actually “LOOK” at the error messages it looks like undionly.kpxe is making it to the target computer so dhcp IS working, but then when iPXE queries for dhcp its getting a conflicting or no {next-server} value. This makes me thing you have / had another images platform on your network at one time.

[monasmith529]

@george1421 Correct. I have been using fog for about a year. This started happening today after updating the latest RC.

[george1421]

@monasmith529 What was the previous version of fog that worked?

This is a physical machine, right?

[monasmith529]

@george1421 1.3.5 RC10

[george1421]

@monasmith529 so upgrading from RC10 to RC12 on the same hardware gives you this error? Because this kind of resembles a spanning tree issue. because once you enter in the tftp server’s IP address you are getting a connection timed out error. Meaning the client couldn’t reach the fog server at 172.18.8.24

I’m questioning specifics around this iPXE kernel since we need to find out if iPXE has been updated since RC10. We are not in fog yet so no other changes in FOG should have an impact at this point.

[monasmith529]

@george1421 Yes…same hardware…same image as previously working on RC10. I do not have spanning tree enabled on the switches in my tech room for imaging.

[george1421]

@monasmith529 Well then I guess we need to ping the @Senior-Developers to see if iPXE was recomplied between RC10 and RC12

Just to confirm this functions the same if you attempt on different, but previously functioning hardware?

[Tom Elliott]

@george1421 No, it wasn’t.

[Tom Elliott]

This, by the way, also appears when you have multiple dhcp servers. I might suggest checking there.

[george1421]

@Tom-Elliott That is where I was heading at first with multiple conflicting responses, but then typing in the ip address of the FOG server, iPXE should have continues and not timed out. That is why I was also thinking about spanning tree being enabled.

The iPXE kernel must be picking up a dhcp IP address, but then it falls down.

[Tom Elliott]

@george1421 TFTP itself is working, but is default.ipxe actually on the server at /tftpboot/

[monasmith529]

@Tom-Elliott Well, I came in this morning and its working fine. I made no network changes over night, so the gremlins must have let loose. Thanks for your quick responses and support!

[Tom Elliott]

I’d put a “watch” on the network then @monasmith529 because this really sounds like something is being a “Rogue” dhcp controller on your network.

[monasmith529]

@Tom-Elliott already done