Hello,

 

I am contacting you regarding an abnormal behavior observed on a Dell Precision T1700 used as a Proxmox VE server.

 

Hardware and software information:
- Model: Dell Precision T1700
- BIOS / Firmware: Dell BIOS A28
- Hypervisor: Proxmox VE
- Network controller: Intel I217-LM
- Linux driver: e1000e
- Network interface: eno1

 

As part of the implementation of a Wake-on-LAN solution, several BIOS settings related to power management and wake-up functions were checked and modified.

After BIOS changes performed in this context, followed by a Save/Apply and Save & Exit sequence, an abnormal behavior was observed: after a normal Proxmox VE shutdown, the server could automatically restart instead of remaining powered off.

 

I performed several tests to determine the cause.

 

1. Proxmox / Linux shutdown analysis

I reviewed the previous boot logs using journalctl.

The shutdown sequence appears normal:
- the shutdown request is received;
- “The system will power off now!”;
- “System is powering down.”;
- shutdown.target is reached;
- final.target is reached;
- systemd-poweroff.service completes;
- poweroff.target is reached;
- file systems and block devices are synchronized.

No explicit reboot command was identified in the analyzed shutdown sequence.

 

2. Wake-on-LAN test

I disabled Wake-on-LAN at the system level using ethtool and changed the state to “Wake-on: d”.

The server was then shut down normally.

The server still restarted automatically.

 

3. Network isolation test

To completely rule out a network wake-up, I physically disconnected the Ethernet cable before shutting the server down.

The server still restarted in this configuration.

This strongly reduces the likelihood that a Magic Packet, Wake-on-LAN, or another network event was directly responsible for the restart.

 

4. BIOS / ACPI investigation

I reviewed the BIOS power-management and wake-up settings.

Wake-on-LAN needs the network controller to remain sufficiently powered in S5. Deep Sleep was therefore disabled.

The system logs also show ACPI support for the platform power states, including S0, S4 and S5.

The kernel logs identify a Dell CBX3 platform with AMI firmware tables and also report:

“ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored”

I understand that this message alone does not prove that the BIOS caused the restart, but it appears relevant to the firmware/ACPI investigation.

 

5. Forced shutdown test

I also performed a forced shutdown by pressing and holding the physical power button, in order to compare the hardware behavior with the normal Proxmox software shutdown.

 

6. Subsequent BIOS operation

I then performed another BIOS modification, explicitly followed by Save/Apply and Save & Exit.

After this operation, the automatic restart was no longer observed during the tests performed. The server remained powered off correctly, and Wake-on-LAN continued to work.

However, I do not consider the issue permanently resolved. The behavior may still be reproducible, and the exact root cause has not been identified.

Based on the tests performed, the issue appears more likely to be related to BIOS/firmware, ACPI, or the transition to the S5 power state than to Proxmox or the network.

 

Could you please confirm:

1. Is this behavior known on Dell Precision T1700 systems running BIOS A28?
2. Are there known issues involving ACPI, S5, or automatic restart after a normal shutdown on this platform?
3. Is there a newer BIOS/firmware version or a corrective update for this behavior?
4. Is there a Dell procedure to fully reset the BIOS power-management/ACPI state?
5. Is there a specific BIOS setting known to cause this behavior after Save/Apply and Save & Exit?
6. Is there a Dell diagnostic method to identify the hardware or ACPI event responsible for a wake-up/restart after shutdown?

 

The situation is currently controlled and the server is operating normally, but I am reporting the issue in order to determine whether it is a known problem and, if possible, obtain a permanent corrective solution.

Network IP addresses have intentionally been omitted from this report and email.

 

Thank you in advance for your analysis as communauty and support ! ;) 

 

Best regards,

Aqlyvornix