@Tom-Elliott
It seems that Machine 17 is the one having the problem with the FOG client—I can’t fix the issue there even by reinstalling it.

Why does the PC have so many network cards?
4b371f64-b608-44df-9f98-69df64096c3b-grafik.png

@Tom-Elliott said in Unauthorized error:

@Jamaal I’m not sure I follow? The database isn’t insecure, it was being able to download the database, or force an upgrade without authorization that was the piece being secured, nothing about the normal day to day operation of the database changed.

Got it, ok thanks for your help today on this.

@Tom-Elliott

I’ve updated my FOG to the stable release and also updated my script with the following three lines, and everything’s fine – the database has been backed up successfully, it’s perfect!

fogApiToken="XYZ=" fogUsrToken="ZYF" curl -ik -X GET "http://$fogServerAddress/fog/system/export" -H "fog-api-token: $fogApiToken" -H "fog-user-token: $fogUsrToken" -o $backupDir/mysql/fog.sql 2>>$backupDir/logs/error.log 1>>$backupDir/logs/progress.log 2>&1

Thanks again for all your hard work on the FOG project!

@JCS-RVK Understood and apologies there’s not a good method around this.

Thanks for at least understanding. I wonder if some of the similar types of issues we’ve seen historically have a similar style of problem that I’m only coming to realize now-a-days.

@Tom-Elliott

Just lyk that I updated the main FOG server to Debian 13 and installed the latest stable release of FOG.

Replication working well now that the OS and FOG server versions match the Storage Node server.

Thanks again

Works well ! 😊

Works well ! 😁

@Tom-Elliott yes, I checked the services, they all running.
systemctl status nfs-server
● nfs-server.service - NFS server and services
Loaded: loaded (/usr/lib/systemd/system/nfs-server.service; enabled; preset: enabled)
Drop-In: /run/systemd/generator/nfs-server.service.d
└─order-with-mounts.conf
Active: active (exited) since Tue 2026-06-30 09:06:50 UTC; 17h ago
Process: 4995 ExecStartPre=/usr/sbin/exportfs -r (code=exited, status=0/SUCCESS)
Process: 4997 ExecStart=/usr/sbin/rpc.nfsd (code=exited, status=0/SUCCESS)
Main PID: 4997 (code=exited, status=0/SUCCESS)
CPU: 10ms

Jun 30 09:06:49 fog-svr.local systemd[1]: Starting nfs-server.service - NFS server and services…
Jun 30 09:06:50 fog-svr.local systemd[1]: Finished nfs-server.service - NFS server and services.

rpcinfo -p 192.168.13.99
192.168.13.99 mountd
rpcinfo -t 192.168.13.99 nfs
rpcin program vers proto port service
100000 4 tcp 111 portmapper
100000 3 tcp 111 portmapper
100000 2 tcp 111 portmapper
100000 4 udp 111 portmapper
100000 3 udp 111 portmapper
100000 2 udp 111 portmapper
100024 1 udp 50174 status
100024 1 tcp 34667 status
100005 1 tcp 20048 mountd
100005 2 tcp 20048 mountd
100005 3 tcp 20048 mountd
100003 3 tcp 2049 nfs
100227 3 tcp 2049 nfs_acl
100021 1 udp 38178 nlockmgr
100021 3 udp 38178 nlockmgr
100021 4 udp 38178 nlockmgr
100021 1 tcp 37227 nlockmgr
100021 3 tcp 37227 nlockmgr
100021 4 tcp 37227 nlockmgr

I deployed the image using a VM on Proxmox and it worked, but with a phisical PC it gives the error like above. All are on the same VLAN.

@Tom-Elliott awesome, thank you!

@Tom-Elliott

Ah man that’s exactly what I needed. Thanks so much sir!

Thanks @Tom-Elliott !

Perhaps we could include a brief note stating that it is possible the database schema may not need updating.

@Nono Edit your /opt/fog/.fogsettings file and replace entries of ‘mysql-client’ with ‘mariadb-client’

I suspect they’ve updated the repo to solely exist for mariadb-client, but you had it before that package switch/change occurred so your fogsettings file is just expecting mysql-client always.

I might also recommend removing the existing mysql-client package just to ensure clean flow, though of course get a backup before any such actions.

@vanfifty1 I feel like this was something you produced on GIthub as well, is this correct?

I’m just wanting to link things. As it’s usually usually me doing all the development now-a-days again, (but thank you AI) I can try to have Claude take a look.

I’m currently working on 26H1 update download at least thru a vm, but just wanted to give a heads up.

@Valer Hi Valer,

You can see my tutorial on using Secure boot with Shim, and my thoughts on what 2.0 means for Secure Boot with FOG here: http://forums.fogproject.org/post/158170

Shoutout “The Minester”

https://github.com/FOGProject/fogproject/issues/685

On the storage node, create a .my.cnf file in root’s home directory:

sudo tee /root/.my.cnf << ‘EOF’
[client]
skip-ssl = true
EOF

MariaDB will try to connect with SSL by default but that’s not enabled on the main FOG server. This is why I had to use the --ssl=FALSE flag when I connected to mysql manually.

The installation succeeded!

@Florent Hi Florent,

I actually have been meaning to look into this some more, but the likely answer is no, or at least, not entirely. The way that support works is, you download a signed iPXE 2.0 binary from iPXE and a copy of their signed shim. That shim is signed with the Microsoft keys and trusts the iPXE signing keys. What this means in practical terms is, all the steps above would still need to occur, it’s just that the signing of the iPXE binary is managed by iPXE, and you don’t need to enroll a key to boot iPXE.

That said, I would imagine this only covers you for booting iPXE, any chainloaded binaries would still need to be signed either with Microsoft’s key or a MOK key you’ve enrolled on the machine. In FOG’s case this means the FOS kernel has to be signed and trusted on the system, in addition to any other binaries (for example memtest, refind) you plan to boot via FOG.

The other likely blocker is the build itself. Naturally, only iPXE can sign binaries that the iPXE Shim will support. Currently the FOG installer actually builds a slightly modified iPXE binary from source. While I’m unsure if these are all that different from the pre-built binaries from 2.0 in terms of support and functionality, it would at the very least need to be changed to instead pull the iPXE 2.0 binaries.

I don’t think any of these are particularly hard to overcome or deal with though. The bottom line is, 2.0 makes it easier, but only to a point. To get real proper Secure Boot support in FOG, they’ll likely need to generate their own signing keys, and start signing at least the FOS kernels (if not iPXE itself) and update FOG to include shim support somehow.

That said, for basic support, I doubt they would need to go the full mile and get a Microsoft approved signing key, I think distributing a certificate/key you can enroll via MokManager and using a pre-existing signed shim (like the iPXE provided one) would more than suffice for most usecases. I’m not sure how difficult it would actually be to implement any of this into FOG, that’s a question for someone who knows PHP and is more familiar with the FOG codebase than I.

Sorry if that’s a bit long winded, it’s not an easy topic to distill. Hope that helps though.

seems to work when copying the directory from another host over to /opt/fog-service and creating a runit file /etc/sv/fog-service/run with content

#!/bin/sh exec 2>&1 echo "Starting FOG Client..." exec mono-service /opt/fog-client/FOGService.exe -d /opt/fog-service -l /opt/fog-service/service.lock

Hello everyone,

I am facing an issue with image capturing after performing an upgrade on my FOG server from 1.5.10 to 1.5.10.1886. Before the update, everything worked fine for me. The images were stored directly on the Synology NAS.

My Setup:

FOG Server: IP 192.168.10.220 (Debian 13) Storage: External Synology NAS with multiple virtual IPs (192.168.109.220 and 192.168.110.220). Storage Configuration: The Synology NAS is configured in FOG web UI as the Master Node for its storage group. The local Default storage node is NOT the master. Clients: Multiple clients on different subnets (e.g., 192.168.109.23 and 192.168.110.23).

The Problem:
The Partclone phase finishes successfully on the client machine. The image files are correctly uploaded via NFS directly to the Synology NAS into the /images/dev/[MAC_ADDRESS] folder.

However, right after Partclone reaches 100%, the task gets stuck in the FOG Web UI (at around 70%), and the client screen shows the following PHP FTP error:

Error returned: Type: 2, File: /var/www/html/fog/lib/fog/fogftp.class.php, Line: 709, Message: ftp_put(/images/dev/[MAC]): Failed to open stream: No such file or directory, Host: 192.168.110.220, Username: foguser

What I have verified:

I tested the FTP connection manually via CMD/PowerShell from a PC using the same foguser credentials. I am able to log in, mkdir, rename, and rmdir inside the /images and /images/dev directories on the NAS without any permission errors. If I move and rename the MAC folder manually inside Synology File Station from /images/dev/[MAC] to /images/[Image_Name], the image works fine. This setup worked flawlessly before the FOG server upgrade. The /images directory is NOT mounted locally on the FOG server itself (and never had to be). Verified FTP username and password on NAS and FOG. It's same.

It seems that fogftp.class.php is incorrectly triggering ftp_put (trying to read a local file from the FOG server) instead of doing a remote ftp_rename directly on the NAS storage node.

Has anyone encountered this bug after a recent upgrade, or is there a specific setting in the new version that I missed?

Thank you for any help.

Storage Node for NAS
Storage Node for NAS.png

Error on PC
U10-PC13.jpg