@george1421 Its DHCP snooping

I guess I found my problem, it was autoexec.ipxe. I don’t remember where, but I read that I can use context of default.ipxe to autoexec.ipxe, because on fresh FOG server, there isn’t the file and every time when client boot, I get:

autoexec.ipxe not found

now i don’t use any autoexec.ipxe file and everything works fine.

So I think we can close, and problem was on Switch configuration and my autoexec.ipxe file.

Thanks a lot!

@george1421 I disabled Snooping on the switch, and now I can register the UEFI All-in-One with UEFI - it works every time.

But when I try to create an image from the FOG web console and capture image, it breaks everything. I get the same error on all devices, BIOS-based All-in-Ones and even VMs on the same host failing with the same error.

Tomorrow I’ll try to find out after what action exactly this happens

@george1421 Got it, Thanks. Next week I’ll try using a different switch — maybe that will help. Right now, there’s only one Cisco switch between server and client, so I’ll try changing the model to see if it makes any difference.

All the Toshiba All-in-One devices are the same model, and I’ve tried several of them. As I mentioned earlier, everything works fine on a different All-in-One model that only supports BIOS - it’s some unbranded device from China.

@george1421 Yep, this happens after the PXE boot menu, when I select “Quick Registration”.

I’m running the latest stable version of FOG 1.5.10.1667
bzImage: 6.12.25
Init: 20250429

The affected machines are Toshiba all-in-ones that support only UEFI, 8Gb RAM, and NMVe from Western Digital

The problem is that it works at first, but then it suddenly stops. In most cases, it works fine right after a fresh FOG installation with the default kernel, and default configuration.

I’ve already tried setting up new FOG servers multiple times when changing kernel (bzImage and Init.xz) didn’t help. Then I install a fresh server again, and it works — but after a day, the issue comes back.I tried almost already everything on internet

I’ve done this about 10 times — it doesn’t work at all. Then on the 11th try it suddenly works, but again only for a day before failing with the same error. Still I didn’t change anyting on server or client. Server I tried use Fedora 38 and Ubuntu 12.8.0.

So it’s really inconsistent and only happens on UEFI clients. BIOS clients most time work fine.

Maybe something wrong with Cisco Switch, cuz couple years ago it works without problem on this Toshiba all-in-ones with Mikrotik Switch, and use use portfast on Cisco Switch port.

I’m facing an issue when PXE booting on PCs with UEFI. The boot fails with the following error:

Kernel panic - not syncing: VFS: Unable to mount root fs on "/dev/ram0" or unknown-block(1,0)
Kernel Offset: disabled
---[ end Kernel panic - not syncing: VFS: Unable to mount root fs on "/dev/ram0" or unknown-block(1,0)
This issue only happens on UEFI clients — legacy BIOS boots work fine.

What I’ve tried:
Tested different DHCP boot file options: ipxe.efi and snponly.efi — same result
Tried both latest and older versions of bzImage and init.xz — no change

The DHCP server runs on the same FOG Project server.
My topology is: ESXi Host → Cisco Switch → Client.
Other PCs with BIOS boot normally - but UEFI clients always fail with the error above.

Any ideas what might be causing this or how to fix it?

Thanks in advance!

@george1421 Thank you so much! I enabled portfast on the switch port, and it works on physical PCs.

Kernel panic - not syncing: VFS: Unable to mount root fs on "/dev/ram0" or unknown-block(1,0)
Kernel Offset: disabled
---[ end Kernel panic - not syncing: VFS: Unable to mount root fs on "/dev/ram0" or unknown-block(1,0)

But sometimes this error appears even if I haven’t changed anything in the configuration and files. Even if I update or roll back the init.xz or bzImage file to a previous version, it doesn’t help. The only thing that seems to work is setting up a completely new FOG server, but after some time, the problem can come back. How I can find right version or how I can solve this problem? Most time it error on PCs with UEFI

Hello,

I have a problem with getting an IP address after the PXE menu. My FOG Project server runs on an ESXi host. I tried creating a VM on the same host, and it works (both UEFI and BIOS), but when I try on a real PC, I get this error. The network topology is: ESXi server → Cisco Switch → Client.

Starting en01 interface and waiting for the link to come up
udhcpc: started, v1.37.0
udhcpc: broadcasting discover
udhcpc: broadcasting discover
udhcpc: broadcasting discover
udhcpc: no lease, failing
Either DHCP failed or we were unable to access http://ip/fog//index.php for connection testing.
No DHCP response on interface en01, skipping it.
Failed to get an IP via DHCP! Tried on interface(s): en01
Please check your network setup and try again!

I tried many ways, like configuring DHCP on my Windows Server and also using the DHCP built into the FOG Project, but every time it works on a VM, but not on a real PC. On the forums, I read that it could be because of the init.xz file. I tried changing the version, but with some versions I get an error about mounting the root fs.

Also I tried run in on Debug Mode, but i get error:

Unable to mount root fs on "/dev/ram0" or unkown-block(1,0)

I tried two different types of PCs, both with UEFI and BIOS modes. I also checked our firewall and I don’t see any traffic between the client and the server, and nothing is being blocked.