Skip to Content
Cloud serversServer management

Server management and recovery

Power and reboot

Power on starts a powered-off machine. Power off cuts virtual power; shut down inside the operating system first to prevent filesystem corruption. Reboot restarts the system. Dashboard status may take a short time to update after a command inside the guest.

SSH and VNC connections

For SSH, use the image’s username, public IP and corresponding private key. VNC creates a short-lived, single-use ticket and provides console access when networking or SSH is broken.

ssh -i ~/.ssh/id_ed25519 root@SERVER_IP

SSH keys and password reset

The dashboard stores only the public key. During creation, the selected key is added to authorized_keys, and password authentication may be disabled. Password reset creates new credentials; change them after signing in.

Networking and access control

  • Firewall: inbound/outbound rules and server attachments; test your administrative SSH rule before restricting access.
  • Private network: private communication between servers in compatible zones; use private IPs for internal traffic.
  • Floating IP: an address you can assign/unassign to move an endpoint; check destination health before changing it.

The provider firewall does not replace the operating system firewall. Keep both layers consistent.

Metrics

CPU, memory, disk and network charts help identify resource pressure. For the precise cause, also inspect top, free -h, df -h, journalctl and service logs inside the guest.

Backups, snapshots and rescue

Enable or disable periodic backups in Recovery. A snapshot is a manual recovery point. Rebuilding from a snapshot replaces the destination disk’s contents. Rescue is used to repair filesystems, networking or credentials; disable it afterward and test normal boot.

Rebuild, resize and delete

  • Rebuild: reinstalls the operating system from an image and erases disk data.
  • Resize: changes resources to a compatible option; check backups and filesystem compatibility first.
  • Delete: releases the machine and removable dependent resources, and cannot be undone.
Do not treat a snapshot or volume on the same server as your only copy. Maintain an independent copy of important data and test restoration.

Maintenance checklist

  • Security patches and scheduled reboots;
  • User, SSH key and listening-port reviews;
  • Disk space and certificate expiry checks;
  • Backup/restore testing;
  • Firewall and network change records;
  • Removal of unused credentials and resources.
IntroductionIntroduction