Thank you for the information. I used a bare metal machine. I started with a VM but then just used a old pc. I tried it with Fedora then Ubuntu but did not have any luck. I did not get any errors when installing with the install script.

I will take a look at the information you gave me and see if I can get anything working.

@collegestjoseph Can you try out this testing kernel and let us know if it works? https://github.com/rluzuriaga/fos/releases/tag/EXP_20250927

@Jchinn412 The answer is it depends on how you have your disk structure setup

Post the results of these two commands.

lsblk
df -h

It maybe easy or harder (but not impossible) depending on how you have the disk created.

@Fog_Newb Yep, it’s as I suspected:

The Line:

Subsystem sftp /usr/lib/openssh/sftp-server

should be changed to:

Subsystem sftp internal-sftp

Then restart ssh services: systemctl restart sshd

Then your Storage Node testing should succeed!

@george1421 I see now and you’re right as my clients are all legacy boot/BIOS boot non UEFI and would not benefit from the client-arch examination.

I’ve already declarations set for each host in my dhcpd.conf file in terms if MAC to IP and so adding another field of filename “some boot loader file” won’t be impossible.

Thanks you for this exercise as I’ve learned some very important things here.

Because I’m not affiliated with FOG in any official capacity and I’ve never actually contributed to it in any way before, I completely avoided making any actual changes to the source. I have noticed that there could be a couple of things that could make the docker implementation a little easier though, specifically around the FOG services themselves. I’m not entirely sure they are completely container friendly, but I really need more hands on this to test it out thoroughly.

Also, I used Debian 13 as the base image for this and it ends up being pretty large (2.5GB). It might be possible to shrink this with alpine-linux but I didn’t want to break a bunch of stuff and then have to spend more time chasing those bugs. Anybody that wants to work that angle, please feel free.

@Frode-Woldsund

I created an image. Give it a go and let me know

https://forums.fogproject.org/topic/18000/unofficial-docker-image-beta

@tchavei2

check out my image! Let me know if it works for you

https://forums.fogproject.org/topic/18000/unofficial-docker-image-beta

My RP address was not pointing to my FOG server 🤦

Once this was adjusted, multicast is functional. Definitely a good reminder to start small.

Hi

In fact, I really want to use those third party package because they do things very well
My snapin configuration are not tied to a specific version of a package, it’s the upstream repo who take care of that

But I have found how to install my install.bat file

Here is the solution

Snapin Name: 7zip Snapin Description: Installation de 7zip Snapin Type: Snapin Pack Template Snapin Pack Template : PowerShell x64 script Snapin Pack File: "%SYSTEMROOT%\sysnative\windowspowershell\v1.0\powershell.exe" # autofilled Snapin Pack Arguments: -ExecutionPolicy Bypass Set-Location "$env:ProgramFiles\FOG\tmp\7zip"; & ".\install.bat" Snapin File (exists): 7-Zip_25.01-1_x64.zip Snapin Enabled : x Snapin Arguments Hidden: Snapin Timeout (seconds): 0 Replicate?: x Reboot after install: Shutdown after install: Snapin Command read-only: "%SYSTEMROOT%\sysnative\windowspowershell\v1.0\powershell.exe" -ExecutionPolicy Bypass Set-Location "$env:ProgramFiles\FOG\tmp\7zip"; & ".\install.bat" # autofilled

It works like a charm

Maybe it should go in FOG documentation as an example

The third party repo I mentioned is there https://gricad-gitlab.univ-grenoble-alpes.fr/legi/soft/trokata/winsoft-main/

LEGI is a French research lab, and the repo is under MIT Licence. I am not the owner, but it’s a great work. I will explain more about that later.

I found the issue!

The udpsender path on my system was “/usr/sbin/udp-sender”

The path that is set within config.class.php is:

define('UDPSENDERPATH', '/usr/local/sbin/udp-sender');

I ran the following and restarted the FOGMulticastManager to resolve:

sudo ln -s /usr/sbin/udp-sender /usr/local/sbin/udp-sender

And we have logs now! Time to begin multicast testing 🥳

@Tom-Elliott thanks for the quick response. No joy though.

I’ve been and tried an Ubuntu live iso with a 6.14 kernel from the Ubuntu Concept Snapdragon Elite branch (https://discourse.ubuntu.com/t/faq-ubuntu-25-04-on-snapdragon-x-elite/61016) and even that doesn’t boot - though it gets as far as telling me it can’t create a device tree. So I guess I’m just going to have to be patient while upstream gets this whole SoC working.

@Tom-Elliott Yes this did work. Thank you for your quick action.

@BrightPipe Can you post the snapin?

This isn’t something FOG is doing, though you already suspect that as well.

So seeing the snapin file, I’m guessing it’s using some method that’s effectively saying “set all the things” and the snapin is set to only set “specific” things.

Since the action is effectively “setting all the things” it’s resetting the other things you didn’t intend.

Seeing the snapin and the actions involved will be most beneficial at least to understand what’s happening. We may not be able to fix the issue but at least we could all better understand the problem and maybe think of ways to troubleshoot and hopefully narrow down whats wrong.

At cursory glance, it feels like the snapin is misconfigured in someway. Thought you did say it worked in Windows 10, but not in Windows 11, so it’s also just as possible there’s some functional difference that needs to be adjusted for in the Windows 11 side of things.

@rhysb92 We need more information.

i ask because I have just built a new VM, and installed dev-branch/stable and it completed and I was indeed able to login with the fog/password combination where i then changed the PW.

So I feel like we’re missing some key information.

@mr-Twister Can you please try this:

http://forums.fogproject.org/post/122702

effectively your boot menu likely needs to be:

params param mac0 ${net0/mac} param arch ${arch} param qihost 1 kernel bzImage loglevel=4 initrd=init.xz root=/dev/ram0 rw ramdisk_size=127000 web=http://${fog-ip}/fog/ consoleblank=0 rootfstype=ext4 mac= ftp=<storageIPHere> storage=<storageIPHere>:/images/ storageip=<storageIPHere> irqpoll chkdsk=0 capone=1 type=down img=<imageNameHere> imgType=n imgPartitionType=all imgid=2 osid=<imageOSIDHere> imgFormat=<imageFormatHere> imgfetch init.xz boot

Replacing as appropriate:
<storageIPHere>
<imageNameHere>
<imageOSIDHere>
<imageFormatHere>

with your relevant information.

Ultimately finding these pieces of data relevant to your specific image will be the hardest part, but once done I believe this will do what you are wanting/hoping.

@Bearr1976 If you use the latest init, the primary drive is supposed to be /dev/sda automatically (technically the first in alphabetical order) so by setting Host Primary Disk, may be forcing it to use that specific drive. In your case, if you’re trying to capture multiple disks, I suspect you would need the Host Primary Disk unset. If there’s something else not working right after unsetting please let me know and I’ll try to see what I can do to replicate the issue.

@Tom-Elliott
Of course, it’s disabled. 😉