There are multiple new problems related to this issue.

First one is buildipxe.sh script and ipxe with commits after 01/14/26. It does not enable USB_HCD_USBIO correctly.:

sed -i 's+//#define USB_HCD_USBIO+#define USB_HCD_USBIO+g' config/usb.h sed -i 's+//#undef USB_KEYBOARD+#define USB_KEYBOARD+g' config/usb.h sed -i 's+//#undef USB_EFI+#undef USB_EFI+g' config/usb.h

but it doesn’t work correctly with latest ipxe source files because of the tabs. It should be:

sed -i 's+//#define USB_HCD_USBIO+#define USB_HCD_USBIO+g' config/usb.h sed -i 's+#undef USB_KEYBOARD+#define USB_KEYBOARD+g' config/usb.h

USB_EFI is already defined in the source, so last sed is not needed.

But after fixing this and compiling, keyboard does not work anyway. There were significant commits to ipxe source on 1/15/26. They changed usb.h along with defaults. I was able to make it work again by going back to 6cccb3bdc00359068c07125258d71ce24db5118a commit, enabling USB_HCD_USBIO in usb.h and disabling “#define USB_CMD” in general.h that FOG uses (without this it won’t compile giving error about multiple definition of `usbio_driver’).

Adding the menu manually with the above settings resolved the issue for me.

@Tom-Elliott I’m having this issue currently running Fog version 1.5.10.1593 and OS version 10.0.19044 Build 19044. I’ve tried all of these steps and a plethora of others but have been unable to successfully capture a final image. Do you have any other ideas to try?

After upgrading to 1.5.10.1754 it works just fine.
Thanks for bug tracking and improvement!

@thezman007 I would say the pcap file you provided is a model of how a proxy dhcp and dhcp server should interact. The first part of the pcap is perfect.

The second part starting at second #19. The client issues a dhcp discover and the dnsmasq answers right away, the client had to issue a second discover request before the main dhcp server @ 2.2 address responded. This pattern is repeated at the end of the pcap (you can see this if you look at the pcap with wireshark).

So this is only me reading the tea leaves but I think there is something up with your main dhcp server because its being slow to respond to dhcp requests. Understand I only can see 25 second pcap but I find it abnormal. When things go sideways (and it probably will) get a pcap of the failure, that’s going to tell us what’s missing.

I’m going to remove your pcap from your post because its not needed now.

@Tom-Elliott Thanks Tom, yes that looks to have sorted it out thankyou…

@Jamaal This problem is solvable but it make take some effort on your part.

Lets start with the basics.

For the DHCP IP zone where your pxe booting clients live, you need to set dhcp options 66 to the IP address of your fog server. And for dhcp options 67 that needs to be snponly.efi or snp.efi. With those settings configured on a MS Windows based dhcp server a pxe booting client should boot. Make sure on your dhcp server that is responding to bootp and dhcp requests. Its been a while since I messed with windows but on the dhcp server there should be a setting of dhcp bootp or both. Select both.

Now lets talk about WDS for a second. A WDS server can use dhcp options 66 and 67 as above, but it can also run a proxy dhcp service that tells the client to ignore the dhcp options and come talk to it for boot information after it gets an IP address for the dhcp server. This maybe called a netboot service or something like that on your WDS server. Its not part of the main WDS service. If this service is still enabled it will override any settings you make in dhcp for pxe booting.

So how do you figure this out to what’s wrong?

The easiest and most complicated issue is to identify what is flying down your network during the pxe booting process. You can do this with wireshark on a witness computer (computer not part of the pxe booting process). This witness computer can either be a ms windows or linux computer, the key is to have wireshark loaded. When you start up a capture use a capture filter of port 67 or port 68 or port 4011 That will limit what wireshark sees to only the dhcp packets. Make sure the witness computer is connected to the same subnet as the pxe booting computer.

Start the packet capture and then attempt to pxe boot the target computer. Continue to capture the packet until the pxe booting computer either reaches the fog iPXE menu or errors out. Then stop the capture.

In the top section you should see the DORA (discover, offer, request, and finally ack/nack) process. The process goes as follows:
Client -> Discovery
Server-> Offer
Client -> Request
Server -> Ack/Nack

In this process you are most interested in the one or more OFFER packets. In a normal network you should only see one OFFER packet. When WDS is involved you will see one OFFER packet from your main dhcp server and a second OFFER packet from your WDS server. If you are seeing the OFFER from your WDS server then you don’t have the proxy-dhcp service disabled, and that is causing your issue. If you are seeing two offer packets from two different dhcp servers, such as a primary / secondary setup make sure both dhcp server are configured to boot from FOG server.

Now what do you do if you only have one OFFER packet and its still not working. This is where you need to select the OFFER packet and then look at the data in the parameters box. There will be the bootp fields of next-server and boot-file these need to be configured for the fog server IP and snp.efi. Then in the dhcp options section options 66 and 67 need to be set correctly. If one or the other sections are not set correctly you will get random machines not booting while others are.

If you can’t figure it out save the packet capture file “be sure you only captured the dhcp process” and up load the file to a file share site and post the link here and one of us will take a look to see what’s wrong. But I think from what I covered here you should be able to figure out what the pxe booting client is being told to do incorrectly.

@Tom-Elliott I can confirm that it works with v1.5.10.1760. 🎉 Thank you very much! 🙇

@The-Dealman Awesome thank you! and we did publish 1754 specifically due to this issue (manually running the automated processes just in case your org is worried at all 🙂 )

@Tom-Elliott Sorry for the lack of clarity in my explanations and the use of screenshots unrelated to my problem.

I have upgrade to the last beta and deploy on registered host works well.
Thank you.

I’ll check if not registered host now appear in Pending Hosts.

I had some problems with AMD computers and found this

https://knowledge.broadcom.com/external/article/269598/firmware-settings-for-ipxe-boot-on-lenov.html

"Resolution
During iPXE troubleshooting it has been determined that the following settings in Firmware/BIOS for Lenovo Thinkpad L15 Gen 3 need to be adjusted to allow for a successful iPXE boot:

Memory protection - Set to ‘disabled’
AMD V™ Technology - Set to ‘disabled’

There also may be certain environments that require further changes. For further information contact Lenovo Support.

Additional DS (Deployment Solution) and GSS (Ghost Solution Suite) files can be found in KB article Updated 64-bit ipxe.efi (ipxe v1.21.1+) binaries

NOTE: The files in the above KB article are not a pointfix but are new version binaries signed by Microsoft"

Leaving it here, because it solved my problem and it sounds like your problem

@alexamore90 I think you’re asking for something like this? https://forums.fogproject.org/topic/9463/fog-postinit-scripts-before-the-magic-begins?_=1768874256297

@alperi The bit if detail you are missing is what the kernel parameters were that was sent to the fog client. From what you posted it appears that the FOG server has all of the bits in the right spots.

In the kernel parameters that are passed to bzImage during boot up it lists where the FOS engine can find the deployment server. I would verify the IP addresses are correct. If everything appears correct with the parameters, we can debug this a bit more by debug deploy and then manually interact with the fos engine from the target PC’s console.

@Tom-Elliott Alrighty. So my FOG server at another location is acting up in the same fashion with the failure to check in.

Before we used it I updated to 1.60-beta.2265 to hopefully negate it but I did not drop the fog table from mysql before doing so.

Are there any logs you want me to pull?

Was not sure which ubuntu version of these (20.04, 22.04, or 24.04) was used as the basis for the FOG release v1.5.10.1734.

Thanks, @FOGBreaker101, for the version you are using.

Some software doesn’t use the latest released version of the OS distro, but often one major version back.

@CanadienITGuy Just for your and anyone’s fyi the autoexec.ipxe... Not Found is not an error. It’s more of an info message than a warning or error.
I actually have tested adding an autoexec.ipxe, even just an empty file to remove that message but even an empty file or a file that is even just a symlink or copy/paste of our normal ipxe/boot menu files causes things to break in the process.
The autoexec.ipxe is meant for adding customization to the ipxe process without needing to re-build the ipxe binary. But my testing with it within the fog workflow was that it’s best to just let that message exist and to see it as it being not found means the process will not be altered from your expected Fog ipxe workflow

@george1421 Thank you for this suggestion. This disabling of “fastboot” worked PERFECTLY! I interrupted the OOBE with Shift + F10 and got the command line window. I shutdown using the “shutdown.exe -s -t 0” command you suggested. Then after booting up, the FOG capture task I had already started took over and it captured 100% as it should have.

Essentially, I had forgotten to disable fastboot like I normally do on EVERY computer I have. So when it “rebooted” it technically wasn’t rebooting the device since fastboot is a misguided absurd hibernation feature Microsoft developed. The error I got was nearly identical to the one in the original post.

Thanks so much for this suggestion George and all!

9d53d15e-eb10-4045-9c90-31f6c156aa7d-image.png

@Infojoe They are one in the same lol and you’re welcome.

@eliaspereira This should be fully fixed in the stable release of 1.5.10.x coming on the 15th of this month and in the dev-branch as of now.
I thought it was already fixed back in September, and it has been working in 1.6 since then but we just got a report of a related issue here https://forums.fogproject.org/topic/18081 which I believe I just fixed.

@tian @DBailey635 @eliaspereira Apologies for missing this post. This was fixed in August-ish of last year, see also:
https://github.com/FOGProject/fogproject/issues/736

I found this searching for a post I wrote about it, as I’m pushing another fix for this for a bug just found in 1.5.10.x

If you update to the latest dev-branch (or what will be stable on the 15th of this month) or give the working-1.6 branch aka 1.6-beta a try, you’ll find the queuing problems fixed.