How to Configure Server Startup Settings After a Power Outage

Configure server startup settings at two levels: set BIOS/UEFI power recovery to “Power On” or “Always On,” then configure required services and applications to launch when the operating system boots. Look for a firmware control named “Restore on AC/Power Loss,” “AC Power Recovery,” or “After Power Loss,” and save your changes before exiting. Choose “Return to Last State” only if recovery should depend on whether the server was running before the outage.

Desktop auto-login is not required when workloads are configured as services or boot-time background processes. Test recovery under controlled conditions by removing and restoring AC power, then confirming that the machine starts, the operating system boots, and required services are reachable. A UPS helps protect hardware and supports orderly shutdown, but its battery may not last through the entire outage.

Set Server Startup Settings for Automatic Recovery

Configure the server’s BIOS/UEFI power recovery control to “Power On” or “Always On.” Verify both layers with a controlled recovery test so restored electricity brings back usable services without desktop auto-login. Automatic recovery depends on the hardware requesting startup, the operating system reaching its normal boot state, and the required workloads becoming available over the network.

Firmware controls whether the machine attempts to start when AC power returns; a Windows, Linux, or application setting is not the first control to configure, and menu placement and wording vary by manufacturer. Save the configuration before exiting, since unsaved changes do not provide reliable recovery behavior. Confirm the selected behavior on the target server.

Choose “Power On” or “Always On” when the server should request startup whenever electricity returns. Alternatively, choose “Return to Last State” when preserving the previous power state is appropriate. Selecting “Off” leaves startup dependent on manual action or a separate wake event. Keep that distinction clear when deciding whether restored power should always trigger a startup attempt or preserve the previous power state.

Applications need their own startup configuration after the operating system boots. Linux workloads can run through an initscript or rc.local, while application products may offer service registration or dedicated startup settings. Leaving auto-login disabled unless strongly needed can reduce exposure at the physical console while preserving background service startup after a reboot.

Testing must check the complete recovery sequence rather than stop when the hardware powers on. Under controlled conditions, remove AC power, restore it, and confirm firmware-driven startup. Next, verify normal operating system boot and check that each required network service and application is reachable.

Include UPS behavior in recovery planning. Recovery remains a hardware, operating system, and application task.

Find and Save the BIOS/UEFI Power Recovery Setting

Enter the server’s BIOS/UEFI during startup, locate its AC power-recovery control, select “Power On” or the appropriate “Always On” option, then save and exit. Firmware controls whether the hardware attempts to start when electricity returns, so make this change there rather than in Windows, Linux, or an application.

Setup access depends on the system, as does the location of the recovery option. Common entry keys include DEL, F1, F2, and F10, but these are alternatives for different systems, not a sequence to press together. No confirmed manufacturer-specific menu path applies across Dell, HPE, Lenovo, Supermicro, and other servers, so avoid treating a particular submenu name as universal.

  1. Start the server and press its system-dependent setup key during startup: DEL, F1, F2, or F10. Use the key appropriate to that machine to enter BIOS/UEFI configuration.
  2. Find the AC-recovery control, including a setting labeled “AFTER Power loss.”
  3. Choose “Always On” where that is the appropriate automatic power-on option available in the firmware.
  4. Save the configuration and exit BIOS/UEFI. Complete the save operation rather than leaving setup with the changes unsaved.

Missing the control in an expected menu does not confirm that the server lacks the feature. Another submenu or a differently worded setting may contain the power-recovery choice. Check for a control describing behavior after AC loss or restoration rather than requiring an exact match to one label above. Naming differences are a reason to inspect the available firmware choices, not to assume that an operating-system setting takes their place.

Read the selected value carefully before saving. Avoid values that do not provide the intended automatic startup. For the specific goal of starting whenever AC returns, use the available “Power On” or appropriate “Always On” choice rather than selecting a value solely because its name mentions recovery.

Saving is part of the configuration, not an optional final step. Unsaved changes do not provide reliable recovery behavior, even if the desired value appeared selected while setup was open. That check establishes the target machine’s actual response instead of relying only on the wording displayed in its firmware menu.

Choose Between Always on and Return to Last State

These choices represent different recovery policies, not interchangeable names for automatic startup. The automatic options differ in their recovery policies. “Off” prevents automatic power-on after AC restoration, so startup requires manual action or a separate wake mechanism.

Setting Behavior When AC Returns Verification Limit
Power On Requests startup whenever AC returns. Names and menu placement vary by manufacturer; confirm behavior on the target machine.
Always On Starts after AC restoration regardless of the previous power state. Verify the actual response rather than relying on the label alone.
Return to Last State Restores the pre-outage power condition. A test beginning with the server on does not verify its response when previously off.
Off Does not automatically start after AC restoration. A separate wake mechanism can still cause startup.

Prior power state is the deciding distinction when comparing the automatic options. Consider a server that was running when electricity failed: both “Always On” and “Return to Last State” can bring it back after AC restoration. By contrast, if the machine was already off, “Always On” requests startup anyway, whereas “Return to Last State” preserves the off condition. That difference determines which policy matches the intended recovery behavior.

Reports that many servers use “last state” by default should not be treated as a universal manufacturer rule. Community discussion describes that behavior as common, but does not establish a default for every server or firmware configuration. Even the reported automatic recovery of PowerEdge R920 and R910 systems does not confirm which exact setting those machines used. An observed restart alone therefore cannot distinguish “Always On” from a matching last-state response.

Verification should account for the condition that separates these policies: whether the server was on or off before AC disappeared. Testing recovery only from a running state can confirm automatic startup, but leaves the previously-off case unconfirmed. For “Return to Last State,” verify both conditions.

Finally, keep the policy’s scope clear: these firmware choices control hardware startup, not application readiness. Neither automatic option by itself confirms that the operating system will finish booting or that required workloads will become reachable.

I like hardware stores when I need nothing, which is how I end up studying labels until a simple choice develops several unhelpful conditions.

Start Server Applications Without Desktop Login

Use the previously described service or boot-time configuration. Keep automatic desktop login disabled unless there is a strong reason to enable it. Background workloads can run independently of a user session, so signing someone in is not the mechanism to rely on for application recovery. Avoiding auto-login also reduces exposure at the physical console while preserving service startup.

On Linux, an initscript or rc.local can start background programs during boot without requiring a user to log in. Treat these as examples of boot-time startup mechanisms, not as universal instructions for every Linux installation. Required workloads should therefore be configured through the startup facilities applicable to the operating system and application, rather than through desktop auto-login.

Vectorworks Site Protection Server provides a Windows-service registration option with a service name. Registering the server this way makes automatic launch available, including where no browser exists on the server. Startup options must be supplied at every launch to remain active, so service registration and application options deserve separate attention. A service registration addresses automatic launch; the selected startup options control details such as which license files the server uses.

For Vectorworks command-line startup, shut down the server software, open Command Prompt on Windows or Terminal on Mac, and navigate to the Vectorworks Site Protection Server folder. Run the startup command with a dash followed by the option name, then press Enter to restart the RLM server. Using rlm.exe -c ABCD1234.lic selects that named license file, while rlm.exe -c licenses restricts license-file use to the specified folder. Neither example is a Windows-service registration command.

ServerProtect for Linux offers startup controls through the Linux Service Configuration utility. Procedures vary by Linux distribution, so use the applicable utility rather than treating one command as universal. Within ServerProtect, the path Administration > Startup Settings > system administration tool displays startup-setting help. That path is documentation navigation, not a command that registers every Linux application for automatic startup.

Tailscale automatic startup needs its own confirmed procedure; no confirmed service name, startup command, Windows setting, Linux service unit, or verification command is available here. Do not substitute the Vectorworks commands or ServerProtect controls for application-specific instructions. Remote administration also requires separate secure access planning: keeping desktop auto-login disabled preserves the distinction between background service availability and access to the physical console.

Coordinate UPS Shutdown and Firmware Recovery

Use a UPS to protect the server and support orderly shutdown, while keeping firmware recovery and application startup configured independently. Battery backup complements those settings rather than replacing them: its charge may run out before utility power returns. Together, these measures address different parts of recovery, from protecting hardware and data during a power interruption to starting the machine and restoring its workloads after AC electricity becomes available again.

Uninterruptible Power Supplies

Uninterruptible Power Supplies

Uninterruptible power supplies provide temporary battery power when mains electricity fails, helping servers, networking equipment, and computers shut down safely or keep running through brief outages. Compare standby, line interactive, and online models, along with power capacity, runtime, outlet types, monitoring software, and replacement battery options.

Check on Amazon

As an Amazon Associate we earn from qualifying purchases.

A UPS helps protect the power supply, disks, and data against outages and voltage strikes. However, that protection does not establish that the server can remain running throughout an outage. Runtime may be insufficient for a long interruption, so continued operation on battery cannot be the only recovery plan. Orderly shutdown and automatic recovery remain complementary goals: shutting down protects the system, while firmware and service settings govern what happens afterward.

For an APC battery backup connected to a Linux server, apcupsd is an option for handling power events. Compatible UPS communication and local configuration are required before it can respond to those events. Scripts can run when power fails, when electricity returns, or when the battery becomes low. Those capabilities provide a way to coordinate power-event handling with shutdown planning, but installing the package alone does not establish a completed recovery configuration.

APC UPS Battery Backups

APC UPS Battery Backups

APC UPS battery backups provide temporary power during outages and help protect connected servers, computers, and network equipment from power interruptions. Models differ in power capacity, outlet count, runtime, waveform type, management ports, and features such as automatic voltage regulation and replaceable batteries.

Check on Amazon

As an Amazon Associate we earn from qualifying purchases.

Keep the responsibilities separate when configuring that coordination. Firmware controls the machine’s startup request after AC restoration; Linux service settings or application startup facilities control the workloads after boot. Meanwhile, apcupsd handles configured UPS events rather than replacing either layer.

An orderly shutdown therefore does not, by itself, confirm that applications will return automatically. Likewise, a firmware power-on setting does not confirm that the UPS can sustain operation until utility electricity returns.

Verification should cover the UPS event configuration alongside the server’s recovery settings, not treat battery backup as proof of service continuity. No confirmed sizing calculation, battery replacement interval, or shutdown threshold is established here, so those values should not be inferred from this guidance. Before considering recovery complete, confirm that the machine powers on after restored AC and completes the checks noted above.

Test the Complete Recovery Sequence

Run a controlled test by removing AC power, restoring it, and checking that the server powers on, boots normally, and makes its required services available without local login. Treat those as separate checks: a powered-on machine has not demonstrated successful workload recovery. Keep the test focused on the complete sequence rather than treating the firmware response as proof that every application has returned to service.

Start by observing what happens when AC electricity returns. With automatic power-on configured, the machine should request startup without someone pressing its power button. Confirm that this actually happens on the target server, rather than assuming the saved setting guarantees the expected behavior. Record the hardware result separately from the operating-system result so that a failure to power on is not confused with a failure later during boot.

Next, assess operating-system recovery. Reaching a physical login screen does not, by itself, mean recovery has failed: server services can run without an interactive user session. Leave desktop auto-login out of the test and check whether the background workloads start as intended. Distinguish an operating system waiting for local credentials from an application that requires someone to log in before it starts.

Check service recovery after boot. Successful operating-system startup is only one part of recovery; the needed workloads must also become available. Verify them individually rather than using one reachable application as evidence that all server functions have recovered.

Compare those results with the hardware and boot observations to identify which part of the sequence needs attention before calling the test successful. A service that remains unavailable still needs attention even when the machine appears ready at its console.

Account for the UPS when interpreting the power-loss test. Its protection may affect the test outcome. If the machine never loses power, continued operation does not confirm firmware-driven startup after AC restoration. Complete the recovery check by confirming all required services remain reachable without local login, not merely that the server has returned to its login screen.

Troubleshoot Missing Settings and Unexpected Starts

Check for differently named firmware recovery controls when a setting appears missing, and inspect enabled wake settings when the server starts unexpectedly. Treat laptop support and Windows boot interruptions as separate issues rather than assuming that one firmware change resolves every recovery failure. Menu names and placement vary by manufacturer, so an unfamiliar label alone does not establish that automatic power recovery is unavailable on the system you are configuring.

Another firmware interface may offer “Always On” or “Return to Last State” as recovery choices, and labels may also appear as “AFTER Power loss.” Absence of the expected wording is not confirmation that the feature is unsupported. Conversely, no universal workaround is confirmed for hardware whose firmware does not expose a usable recovery control; verify actual behavior on the target machine.

Laptops warrant particular caution because firmware support cannot be assumed simply from an AC connection or a UPS. A reported ThinkPad case did not restart automatically despite BIOS changes and a UPS, illustrating why configuration alone is not proof of recovery. Separately, a Windows 10 tablet user asked whether automatic startup when power was supplied could be configured. Neither case establishes behavior for every laptop or tablet.

Unexpected starts call for checking both AC-recovery behavior and enabled wake settings. Wake on LAN and wake options for USB, Serial, Parallel, or PS2 are relevant inspection points because enabled wake functions can cause systems to start at unexpected times. Distinguish those triggers from startup following restored AC power before changing the recovery policy. Otherwise, a change intended to correct an unexplained wake may leave the actual trigger unaddressed.

Windows 7 and Windows Vista introduce a separate boot-stage concern: after an improper shutdown or power failure, Startup Repair can launch instead of normal startup. Firmware-driven power-on therefore does not establish that Windows completed its boot. Guidance for that legacy behavior should not be extended to Windows 10, Windows 11, or Windows Server: equivalent behavior and supported mitigation steps remain unconfirmed.

Frequently Asked Questions

What Is the Difference Between Always on and Return to Last State?

“Always On” makes the machine start after AC power returns. “Return to Last State” restores the previous power condition, so a server that was on can restart, while one that was off may remain off. Verify the selected behavior on the target system.

Can Server Applications Start While the Computer Remains at the Login Screen?

Server applications can start as services or boot-time background processes without an interactive login. Linux can use an initscript or rc.local, while application products may offer their own service-registration or startup settings. Desktop auto-login is not required for these workloads.

Does a UPS Replace the Need for Automatic Power Recovery?

A UPS helps protect hardware and can support an orderly shutdown, but its battery may run out before utility power returns. Firmware recovery settings and service auto-start are still needed to restore operation automatically. Test the complete recovery sequence rather than relying on battery backup alone.

Why Might a Laptop Fail to Restart When AC Power Returns?

Beyond that, a reported Windows 7 ThinkPad used as a remote server would not restart without its physical power button, despite BIOS changes and an attached UPS. No confirmed cause was established for that behavior. Perform a controlled test by unplugging and reconnecting the power cord, then checking whether the laptop starts automatically.

Which Wake Settings Should I Check if a Server Starts Unexpectedly?

Check Wake on LAN and wake settings for USB, Serial, Parallel, and PS2 connections. Enabled wake settings can cause unexpected starts. Inspect the AC-recovery setting as well to distinguish power-restoration behavior from other wake events.

Automatic recovery requires both firmware power-on behavior and workloads configured to start during boot. Set the BIOS/UEFI recovery control to “Power On” or “Always On” when startup should follow every AC restoration, then save and exit. Configure required applications accordingly.

Under controlled conditions, remove and restore AC power, verify normal OS boot, and confirm that each required network service and application is reachable.

References

Sources read in September 2026.