While it's on, the server can't be deleted or reinstalled from the console, the API or an AI assistant. Turn it off on the Options tab when you do want to.
It doesn't stop a cancellation for an unpaid invoice. If that happens, we email you first.
This adds a watchdog to your server. Turn it on, then restart the server from the console. The system inside has to keep telling the watchdog it's alive; if it stops, the server is restarted.
Set it up once inside the server, as root.
Ubuntu: add the watchdog driver.
# apt install linux-modules-extra-$(uname -r)
Debian: switch to the standard kernel, then restart the server.
# apt install linux-image-amd64
Ubuntu and Debian: load the driver at every start.
# cat > /etc/systemd/system/watchdog-driver.service <<'END' [Unit] Description=Load the watchdog driver DefaultDependencies=no Before=sysinit.target [Service] Type=oneshot ExecStart=/sbin/modprobe i6300esb [Install] WantedBy=sysinit.target END # systemctl enable --now watchdog-driver.service
Every system (AlmaLinux and Rocky Linux load the driver by themselves): tell systemd to use it.
# mkdir -p /etc/systemd/system.conf.d # printf '[Manager]\nRuntimeWatchdogSec=30s\n' > /etc/systemd/system.conf.d/watchdog.conf # systemctl daemon-reexec
Check it: systemctl show -p RuntimeWatchdogUSec should say 30s, and wdctl should show the i6300ESB timer.
On Ubuntu, run the first command again after a kernel update, before restarting.
Lets you run your own virtual machines inside the server. Turn it on, then restart the server from the console. Check it inside the server:
# grep -c -E 'vmx|svm' /proc/cpuinfo # ls /dev/kvm
If you move the server to another location, it stays on only where that location offers it.
It works even when the network inside the server is down, and it shows the boot messages.
On by default: your server starts again by itself after maintenance on our side. Turn it off if you'd rather it stays off until you start it.