MattC

Thank you - resolved (The GT730 is just for server to boot as its a workstation MB without IGPU). For those searching the forum, as reference for RTX5060 TI; Unraid had loaded nouveau driver (unsure of cause as not part of Unraid). Execute theses commands from an Unraid Terminal and reboot afterwards: Disabled nouveau driver mkdir -p /boot/config/modprobe.d echo "blacklist nouveau" > /boot/config/modprobe.d/nouveau.conf ASUS Workstation/ Threadipper AMI BIOS settings to allow Above 4G encoding Advanced settings -> PCI Subsystem settings -> Re-size BAR Support - switch to Enabled Need to have CSM disabled (Go to BOOT section and set CSM (Compatibility Support Module) to disabled This requires UNRAID to be in UEFI mode (On WebUI click Flash device and enable "Permit UEFI boot mode" checkbox)

Thank you for the suggestion. Disabled, and confirmed in drivers status - strange nouveau was present, because it's a new USB image after I transferred disks to a new machine... Rebooted but same issue with nvidia driver / detection Any other suggestions appreciated tower-diagnostics-20260116-1751.zip

[Edit, added syslog extract] I have 2x NVIDIA GPUs intalled, neither are IO mapped, but Nvidia cannot connect. The GT730 is in first PCIe slot so that 5060 is fully available for AI workloads in Docker. The server is booted into UEFI (not CSM) Nvidia Info: Nvidia Driver Version: 580.95.05 Open Source Kernel Module: Yes Installed GPU(s): NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running. IOMMU group 29: [10de:1287] 01:00.0 VGA compatible controller: NVIDIA Corporation GK208B [GeForce GT 730] (rev a1) [10de:0e0f] 01:00.1 Audio device: NVIDIA Corporation GK208 HDMI/DP Audio Controller (rev a1) IOMMU group 38: [10de:2d04] 61:00.0 VGA compatible controller: NVIDIA Corporation GB206 [GeForce RTX 5060 Ti] (rev a1) [10de:22eb] 61:00.1 Audio device: NVIDIA Corporation GB206 High Definition Audio Controller (rev a1) Jan 16 06:12:47 Tower kernel: nvidia 0000:01:00.0: probe with driver nvidia failed with error -1 Jan 16 06:12:47 Tower kernel: NVRM: This PCI I/O region assigned to your NVIDIA device is invalid: Jan 16 06:12:47 Tower kernel: nvidia 0000:61:00.0: probe with driver nvidia failed with error -1 Jan 16 06:12:47 Tower kernel: NVRM: The NVIDIA probe routine failed for 2 device(s). Jan 16 06:12:47 Tower kernel: NVRM: None of the NVIDIA devices were initialized. Jan 16 06:12:47 Tower kernel: nvidia-nvlink: Unregistered Nvlink Core, major device number 239 Any suggestions appreciated tower-diagnostics-20260116-1412.zip

Thanks... After trying various combinations, just having the DEVICES container variable set (not forcing IP or airscan entry) worked. Having more than one variable set pointing to the same scanner resulted in error or mixed results in finding the scanner.

The web interface shows "[object Object]" in the Device drop-down selector, and there's a pop-up error alert which says "Found no devices", but if I drop to Docker console and run scanimage -L, then sane correctly identifies my network attached scanner (both escl and airscan entries). I tried forcing the IP address and airscan entry through the variables, as well as clearing cache, and clearing devices + reload through web interface, but same result. As scanimage finds the scanner, it could it be an issue with the web interface - I'm guessing(!!!) that it doesn't play nice with a web scanner which serves over 443 but does not present a valid certificate....? scanimage -L device `escl:https://192.168.2.124:443' is a OKI MC363-01C095 flatbed scanner device `airscan:e0:OKI-MC363-01C095' is a eSCL OKI-MC363-01C095 ip=192.168.2.124

Scouring various internet sources (which I unfortunately I didn't record, but generally Linux and FreeBSD sites) I managed to get Samba working consistently for multiple flavours of Windows 10 and 11 by setting auth and protocol constraints. This is my full Samba extra config - I have unassigned devices plugin installed, you can skip that section if you do not also have the plugin installed. #unassigned_devices_start #Unassigned devices share includes include = /tmp/unassigned.devices/smb-settings.conf #unassigned_devices_end #smb local master configuration [global] domain master = yes preferred master = yes os level = 255 ntlm auth = yes client ntlmv2 auth = yes #Uncomment for per-host logging #log file = /var/log/samba/%m.log #max log size = 10000 #log level = 4 #max SMB2 ensures Win10 can connect by preventing upgrade to unsupported SM3.x version client min protocol = SMB2 client max protocol = SMB2 #exclude_Apple_DS_files veto files = /._*/.DS_Store/ Note that adding "ntlm auth" enables NTLMv1 password authentication. This has a vulnerability and should be avoided, but it does provide a fallback if you have some device that doesnt support NTLMv2. To ensure your Windows machine is using NLTMv2, check for the existence of "LmCompatibilityLevel" under "Computer\HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa" . If it is present, it will force windows to use NTLMv1. You should delete this entry, then reboot, and Windows will then revert to using the default NTLMv2. Constraining max protocol to SMB2 prevents Unraid trying to use versions of SMB3 that Windows doesn't support very well - Windows will drop upgrade requests from SMB2 to SMB3.1.1 connections even though it announces them as being supported.

On WIndows 10, after days of fighting Samba, I've simply given up on SMB shares and switched to NFS under windows. I have Win 10 Pro, don't know if it's available in Win 10 Home..... Search for "Turn Windows features on or off", scroll down to "Services for NFS" and expand, select the sub-item "Client for NFS". After searching for the feature it will install, no reboot required. The servers dont show up in the "Network" area of file explorer, but if I enter \\tower.local\ (or \\<server-ip-address>\) then all of the NFS shares are displayed, and I can map network drives or pin to the Quick Access section just like Samba shares.