Free tools Windows power users keep installed
One-click scans. No signup required.
If sudo shutdown -h now has already run, your VPS is powered off and cannot be started through SSH. Open your hosting provider’s control panel, select the stopped instance, and choose Start or Power on. Use the provider’s browser or serial console to watch the boot process, then reconnect with SSH and verify services, mounts, and logs. For a reboot while you still have a live shell, use sudo shutdown -r now or sudo systemctl reboot.
Antihost offers Linux VPS plans with full root access, unlimited traffic on a 1 Gbps port and a 14-day free trial.
What happened when you ran shutdown -h now?
The command selected an immediate poweroff, not a reboot. In the shutdown syntax documented for Ubuntu and systemd-based Linux, now is an alias for +0, meaning the action starts immediately. The -h option selects halt/poweroff behavior unless an explicit halt or poweroff mode changes that choice.
As an Amazon Associate I earn from qualifying purchases.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →After the operating system completes the shutdown, the virtual machine is no longer running. Its network interface, SSH daemon, and applications are unavailable, so another SSH command cannot turn it back on. Starting a powered-off VPS requires an action from the provider’s control plane, management API, or an out-of-band console.
Recovery procedure: turn the VPS back on
- Wait briefly, then test SSH once. Shutdown may still be finishing. A single connection attempt distinguishes a delayed shutdown from a completed poweroff; repeated attempts do not start a stopped guest.
- Open the VPS provider dashboard. Sign in through the provider’s normal web control panel and open the page for the affected instance.
- Check the reported power state. If the panel says Stopped, Powered off, or similar, use Start or Power on. Those controls are intended to boot an instance that is not running.
- Use Reboot only when appropriate. If the panel still reports the VPS as running, or the guest appears stuck during shutdown, choose the provider’s documented Reboot or Reset action. Do not assume every button labeled “restart” performs a graceful operating-system reboot.
- Open an out-of-band console when available. A browser, HTML, serial, or instance console can show firmware, bootloader, filesystem checks, and service-start errors even when SSH is unavailable.
- Reconnect after the system is reachable. Once the console or dashboard shows that the guest has booted, connect with SSH and perform the post-boot checks below.
- Escalate failed control-plane actions. If Start, Power on, or Reboot returns an error, use the provider’s documented API or support channel. Include the instance identifier, the displayed power state, the time of the failed action, and any console output.
Choosing Start, Reboot, Reset, or an API operation
Provider controls operate outside the guest operating system, so their names and effects differ. The safest choice depends on the power state shown by the panel and whether the guest is responding.
| Panel state or symptom | Preferred action | What to verify |
|---|---|---|
| Stopped or powered off | Start or Power on | The instance transitions to running and produces boot output in the console. |
| Running and responsive | Use an in-guest graceful reboot when possible | Services stop cleanly and SSH returns after boot. |
| Running but shutdown appears stuck | Provider Reboot or Reset, following its documentation | Whether the operation is graceful or a hard power cycle. |
| Dashboard action fails | Provider API or support channel | The exact error, current power state, and console output. |
Some platforms expose separate instance-console and reboot management functions. Jetstream2 documents console access and instance reboot management, while VPS.org documents restart and power operations through its API. These examples illustrate why you should read the specific provider’s semantics: one provider’s “restart” may mean a graceful reboot, while another defines it as stopping the virtual machine and starting it again.
If the shutdown has not completed yet
A shutdown scheduled for a future time can be canceled from a still-live shell with:
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutesudo shutdown -cThat cancellation is useful for commands such as shutdown -h +10, where the ten-minute delay leaves time to change your mind. It may be too late after shutdown -h now, because +0 begins the operation immediately. If the shell is still accepting commands, cancel first and then choose the correct reboot command.
The correct commands for future maintenance
| Intent | Command | Result and caution |
|---|---|---|
| Immediate poweroff | sudo shutdown -h now |
Requests an immediate halt/poweroff. The guest will need a provider-side start action to run again. |
| Immediate graceful reboot | sudo shutdown -r now |
Selects reboot instead of poweroff and allows the operating system to stop services and unmount filesystems through its normal shutdown path. |
| Reboot through systemd | sudo systemctl reboot |
Asks systemd to perform a normal reboot from a live shell. |
| Forced reboot | sudo reboot -f or the equivalent forced systemd form |
Bypasses normal init handling. Debian warns that filesystems may not be properly unmounted, so use this only when graceful reboot methods fail. |
For routine maintenance, use shutdown -r now or systemctl reboot. Reserve a forced reboot for a guest that is genuinely unresponsive and where the provider’s console or support guidance indicates that a hard recovery is necessary.
Why SSH cannot restart a powered-off VPS
SSH is a network service running inside the guest. Powering off stops the kernel, network stack, and SSH daemon together. An SSH client therefore has no execution environment in which to run reboot, systemctl, or any other command. Authentication failures, connection timeouts, and “connection refused” messages all have the same practical implication in this situation: use the provider’s management plane rather than trying different SSH options.
The provider control plane communicates with the virtualization host, not with a running shell inside the VPS. That is why Start and Power on remain available after an intentional guest shutdown.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Using the browser or serial console
A console is valuable when the dashboard says the instance is running but SSH does not return. Open it before issuing repeated power operations and watch for:
- Kernel or bootloader messages indicating that the guest is actually starting.
- Filesystem checks or emergency-mode prompts that require a response.
- Network configuration errors that prevent SSH even though the operating system booted.
- Service failures that occur after the network comes up.
If the console shows a login prompt, the operating system is up and the problem is likely limited to networking, SSH configuration, or firewall rules rather than power state. If it remains blank or frozen while the control panel reports “running,” follow the provider’s documented reset procedure instead of assuming that another in-guest command can help.
Post-boot checks after SSH returns
Do not stop at the first successful login. A power transition can leave an important service, mount, or scheduled task in a different state. Run checks appropriate to your distribution and workload:
- Confirm the current boot. Check the system’s uptime and current time so you can distinguish the new boot from an old terminal session.
- Check failed systemd units. Run
systemctl --failed. Investigate any unit listed as failed before declaring the VPS healthy. - Verify disks and mounts. Use
lsblk -fto see attached filesystems andfindmntto confirm that required mount points are present. - Check critical applications. Query each workload’s service status with
systemctl status SERVICE_NAME, replacingSERVICE_NAMEwith the actual unit used by your application. - Review the previous boot’s warnings.
journalctl -b -1 -p warning..alertshows warning and higher-priority messages from the boot before the current one when that journal is available. - Test externally. Verify the ports, web endpoints, monitoring checks, and scheduled jobs that users depend on, not only the SSH connection.
If a mount is missing, do not immediately recreate data or edit /etc/fstab. First compare the device name and filesystem identifier shown by lsblk -f with the configured mount entry, then use the console or provider documentation for recovery if the disk itself is absent.
When the provider reports a contradictory state
The panel says “stopped,” but Start does nothing
Wait for the operation result rather than clicking Start repeatedly. Capture the error message and instance state, then try the provider’s documented power API if one exists. A failed start can be caused by a control-plane issue that only support staff can clear; additional clicks do not provide more information.
The panel says “running,” but SSH times out
Open the browser or serial console. If boot output is progressing, wait for the network and SSH service to finish starting. If the console is frozen, use the provider’s documented Reboot or Reset operation and record that it may be a hard stop/start cycle.
The VPS boots but an application is unavailable
Use systemctl --failed, the individual service’s status command, and the previous-boot journal query. Check mounts before restarting an application that expects data on a separate volume. A successful VPS boot does not guarantee that every dependent service started successfully.
You have no console or API access
The provider’s support channel is the remaining recovery path. Supply the instance name or identifier, the approximate time you ran the shutdown command, the dashboard’s current state, and the exact result of any Start or Reboot action. This gives support enough context to distinguish a powered-off guest from a host-side problem.
Graceful reboot versus a hard stop/start cycle
A graceful reboot lets the guest operating system signal services, flush buffered data, unmount filesystems, and then restart. A hard stop/start cycle is controlled by the virtualization host and may remove power from the guest before those steps complete. Provider APIs sometimes call the latter “restart,” so the label alone is not a safety guarantee.
Before selecting a provider-side reset, check whether its documentation specifies graceful reboot, stop-then-start behavior, or a forced reset. Prefer the in-guest commands while the shell is healthy; use the control-plane operation when the guest is already off or no longer responds.
Preventing an accidental poweroff
- Read the option before pressing Enter:
-hselects halt/poweroff, while-rselects reboot. - Use
sudo systemctl rebootwhen you want an unambiguous systemd reboot command. - For delayed maintenance, schedule a future action and keep
sudo shutdown -cavailable until the window begins. - Confirm that you are connected to the intended VPS before running any power command, especially when several terminal tabs use similar hostnames.
- Keep provider dashboard credentials, API documentation, and console access available before maintenance starts; SSH alone cannot recover a powered-off guest.
The key distinction is simple: shutdown -h now ends the guest’s power state, whereas shutdown -r now ends the current session and starts the operating system again. Once the first command has completed, recovery begins outside Linux with the provider’s Start or Power on control.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
