FOG Could not start download: Invalid Argument

IndustrialTech87

@george1421 I had the same problem a while back on FOG 1.4 and updating the kernel fixed the issue.

HowToGravity

@george1421
Only these new dells are throwing this error?
Yes only the new Dells are having this issue. I just tested with another known working Dell and it is working just fine.

Do you have any custom fog iPXE menu items that you added?
No not that i am aware of everything is default if not close to default. The only thing i changed is the FOG menu time out.

Beyond these questions, I would have you update to the FOS Linux kernel 5.6.x or later. You can do that through fog configuration -> kernel update. You will want the 5.6.x series of kernels to get the latest hardware support.
I believe i have done that, but i did it the manual way because the computer is not attach to the main network. So i downloaded the init.xz, init_32.xz, bzImage and bzImage32 onto a USB then transferred it over. The procedure i used is similar to the steps outlined at the bottom of this wiki https://wiki.fogproject.org/wiki/index.php?title=Kernel_Update.

I doubt the files got corrupted during the transfer because i tested FOG again with a known working laptop and everything appears to be working as intended. But I can repeat the process again to see if that helps.

Is there any way i can check to see if I have the most up to date kernel?

george1421

@HowToGravity What confuses me is these two errors

Could not start download: Invalid argument (http://ipxe.org/1c25e082)
Could not boot: Invalid argument (http://ipxe.org/1c25e082)

This might indicate the syntax of the fog iPXE menu isn’t correct. We’ve seen this where a custom ipxe menu isn’t crafted well. This issue is with ipxe and not with FOS Linux (the OS that boots to capture and deploy images).

If you manually copied over the boot files then navigate to /var/www/html/fog/service/ipxe directory and key in file bzImage the returned kernel image needs to be 5.6.<something>, if its 4.19.<something> then you don’t have the latest kernel. You will need the 5.5.x or later to get the latest hardware drivers. The linux kernel developers are not back porting the new drivers to the 4.19.x series of kernels.

HowToGravity

@george1421 Okay so I file bzImage and it comes up as version 4.19.123. So it does not look like i got the correct bzImage version.

Do you know the link to manually download the updated bzImage & init.xz?

I tried this link but it gives me version 4.19.123. https://fogproject.org/kernels/bzImage

george1421

@HowToGravity At that url https://fogproject.org/kernels/

You want to download Kernel.TomElliott.5.6.18.64 and save as bzImage

and then Kernel.TomElliott.5.6.18.32 as bzImage32

then move those files to the proper locaiton

HowToGravity

@george1421 Unfortunately the Kernel update to 5.6.18 (64 & 32) did not fix the issue. I was successful at capturing and reimaging a known working computer with the new bzImage version, so I am certain that everything was correct.

Here is a link to images of various BIOS settings that can be changed https://imgur.com/a/1loFntW. Please let me know if any of them are not set to the proper value.

george1421

@HowToGravity Ok the kernel update is one thing that is needed, but you are still getting these error messages?

Could not start download: Invalid argument (http://ipxe.org/1c25e082)
Could not boot: Invalid argument (http://ipxe.org/1c25e082)

What exactly did you select to get that error?

FWIW: This error is coming before bzImage loads.

HowToGravity

@george1421
So here are the steps i am taking.

Power on the computer
F12 to get boot option menu
Select the option “UEFI USB LAN1 IPv4 (MAC address)”
The computer proceeds to boot from the network and a few normal looking screens flashes by.
Then i get the following screen with same error. https://imgur.com/y1VPGuO

george1421

@HowToGravity Is this computer currently in an imaging mode? Like it was registered then you picked deploy or capture. When you are booting after scheduling the task this is the error you get?

If so I want you to go to a browser and key in the following

http://192.168.20.2/fog/service/ipxe/boot.php?mac=<mac_of_target_computer>

Then post the output.

HowToGravity

@george1421 Is this computer currently in an imaging mode? Like it was registered then you picked deploy or capture. When you are booting after scheduling the task this is the error you get?
The computer it self is not registered yet with FOG, i still have to get it registered. But from looking at the log, I believe the USB to Ethernet adapter MAC address is registered but under a different computer. So no task has been schedule yet for this computer. My intentions were to boot to FOG then perform a full host inventory and registration.

I am not sure if adding the “-” (hyphen) in the MAC address into the URL makes any difference but I will add both of the outputs below.

without hyphen.txt
with hyphen.txt

*Edit: There are no task queued on the server

george1421

@HowToGravity OK I think we are narrowing in on something.

I’m seeing https in the menu file. Do you have this fog server configured for https services?

This is in the menu

https://192.168.20.2/fog/service/ipxe/bg.png

If you do that from a windows browser it should download the background image for fog.

HowToGravity

@george1421 Not to my knowledge I do not have HTTPS setup, I am the only person who access the FOG server. I followed the link and it does lead me to the background image for FOG, but it is unsecured.

*edit: Interesting HTTPS port is active. I don’t believe I opened it, FOG is the only thing i am using this server for.

george1421

@HowToGravity Setting up https would have been a boot time questions I think.

On the fog server linux console lets see this commands.

netstat -an | grep :80

and

netstat -an | grep :443

What that will tell us if something is listening on http (port 80) or https (port 443). I’m trying to understand this error. The ipxe menu should NOT be throwing this error at this point. I just want to make sure I’m chasing the correct ghost.

HowToGravity

@george1421
netstat -an | grep :80
tcp6 0 0 :::80 :::* LISTEN

netstat -an | grep :443
:::* LISTEN
192.168.20.3:56931 TIME_WAIT
192.168.20.3:56922 TIME_WAIT
192.168.20.3:56923 TIME_WAIT
192.168.20.3:56910 TIME_WAIT
192.168.20.3:56921 TIME_WAIT
192.168.20.3:56933 TIME_WAIT
192.168.20.3:56932 TIME_WAIT
192.168.20.3:56920 TIME_WAIT
192.168.20.3:56930 TIME_WAIT
192.168.20.3:56935 ESTABLISHED
192.168.20.3:56928 TIME_WAIT
192.168.20.3:56929 TIME_WAIT
192.168.20.3:56912 TIME_WAIT
192.168.20.3:56919 TIME_WAIT
192.168.20.3:56918 TIME_WAIT
192.168.20.3:56908 TIME_WAIT
192.168.20.3:56917 TIME_WAIT
192.168.20.3:56927 TIME_WAIT
192.168.20.3:56909 TIME_WAIT
192.168.20.3:56911 TIME_WAIT
192.168.20.3:56913 TIME_WAIT
192.168.20.3:56916 TIME_WAIT
192.168.20.3:56926 TIME_WAIT
192.168.20.3:56934 TIME_WAIT
192.168.20.3:56925 TIME_WAIT

george1421

@HowToGravity Ok you surely have https configured on your server. So that is not it.

Just to be clear here, this problem only happens on this specific model of computer? All other computes work correct, right?

If that assumption is correct, do you have a cheap/dumb/unmanaged network switch?

HowToGravity

@george1421

Just to be clear here, this problem only happens on this specific model of computer? All other computes work correct, right?
Yes, since updating the bzImage this morning I was successful at capturing and deploying an Optiplex 5040 image.

If that assumption is correct, do you have a cheap/dumb/unmanaged network switch?
Yes, exact model is D-link DGS-1008G

george1421

@HowToGravity Well I didn’t get s solid answer from the web site or the manual.

Here is what I’m thinking

Delete the host LIB1723561 (what the nic is registered already as).
It might be a spanning tree protocol issue, where one of the fast spanning tree protocols are not being used.

Searching the manual I could only find 1 reference to spanning tree but not a way to adjust it. If the network port where this computer is being imaged from works with other computer hardware its probably not a spanning tree issue.

This one has me a bit stumped at the moment.

What is the content of the default.ipxe file on the fog server it will be in /tftpboot/default.ipxe. That is the last file it downloads from the fog server’s disk. After that things go sideways.

If you get that message about hitting s for the iPXE shell lets do that. There has to be something I’m missing here because other computers can get into the FOG menu just not this computer

HowToGravity

@george1421
So i deleted the host LIB1723561 and i am still getting the same error as before.

This is the contents of /tftpboot/default.ipxe

#!ipxe
cpuid --ext 29 && set arch x86_64 || set arch ${buildarch}
params
param mac0 ${net0/mac}
param arch ${arch}
param platform ${platform}
param product ${product}
param manufacturer ${product}
param ipxever ${version}
param filename ${filename}
param sysuuid ${uuid}
isset ${net1/mac} && param mac1 ${net1/mac} || goto bootme
isset ${net2/mac} && param mac2 ${net2/mac} || goto bootme
:bootme
chain https://192.168.20.2/fog/service/ipxe/boot.php##params

Any specific commands I should enter in the iPXE shell after hitting s

george1421

@HowToGravity said in FOG Could not start download: Invalid Argument:

chain https://192.168.20.2/fog/service/ipxe/boot.php

If you key in the above do you still get an error? I’m going to say yes, but I’m wondering if there are any more helpful error messages.

HowToGravity

@george1421
I only get the error:
Could not start download: Invalid argument (http://ipxe.org/1c25e082)

FOG Could not start download: Invalid Argument

226

11.6k

17.1k

154.6k