Resources > Resource Management

Reinstalling the OS

Updated

Reinstall wipes the server's disk and lays down a fresh operating system. It's the right tool for a clean start, switching to a different distribution, or recovering from an install that's gone off the rails. Both VPS-class and dedicated servers support it - the customer-side flow is almost identical for both.

Before you start

  • Everything on the disk is erased. Back up anything you want to keep before you start. There's no undo.
  • Have an SSH key added to your account, or be ready to set a root / initial password in the form. You need at least one - otherwise the new install would have no way to let you in.
  • The VPS or server must not be suspended. If the status pill shows suspended, contact Support before reinstalling.
  • On dedicated servers, the location's remote-management agent must be online. The Reinstall page warns you if it isn't and disables the button.
  • Reinstalls take a few seconds for VPS (we deploy from cached images) and 10–30 minutes for dedicated (PXE boot, partition, install, reboot).

Template customizations

Our VPS images are the standard upstream cloud-init builds for each distribution (Ubuntu, Debian, Rocky, etc.), with one addition: the SSH login greeting (/etc/motd) includes Synteq branding. No other files, packages, or configs are modified from the stock image - this is the only customization applied.

1. Open the Reinstall flow

  1. In the sidebar, click Cloud (or GPU Cloud / VDS / Metal).
CleanShot 2026-06-13 at 20.18.51@2x.png
  1. Click the row for the server you want to reinstall. The detail page opens at /vps/{id} for VPS-class or /metal/{id} for dedicated.
CleanShot 2026-06-13 at 20.19.13@2x.png
  1. There are two ways to start a reinstall - pick whichever is closer to your cursor:
  • From the toolbar in the top-right, click the Reinstall icon (a hard-drive icon, far right of the power controls). [1]
  • From the tab row - near the top of the page, click the Reinstall tab. [2]
CleanShot 2026-06-13 at 20.20.27@2x.png

2. Pick the operating system

The page groups available images by OS family - Ubuntu, Debian, Rocky, AlmaLinux, Fedora, CentOS, Arch, and Windows (where it's supported on your product).

  1. Click a family logo (for example, the Ubuntu logo). The tile expands to show available versions as small chips: 22.04, 24.04, and so on.
  2. Click the version chip you want. The chip highlights to confirm.
CleanShot 2026-06-13 at 20.21.11@2x.png

3. Set credentials

Scroll down to the credentials section. You need at least one of these - without one, you'd lock yourself out of the new install.

  1. SSH key: tick the checkbox next to any of your saved keys. You can tick more than one. If you don't see the key you want, click Add new key to paste a public key inline; it's saved to your account.
  2. Password: type an initial password (minimum 8 characters), then retype it in the Confirm password field. The two must match before the Reinstall button enables.
CleanShot 2026-06-13 at 20.21.41@2x.png

4. Confirm and start

Scroll to the bottom of the page. There's a final action bar with a confirmation checkbox and the Reinstall button.

  1. Read the warning text. Tick the I understand all data will be wiped checkbox.
CleanShot 2026-06-13 at 20.22.19@2x.png
  1. Click Reinstall.
CleanShot 2026-06-13 at 20.22.44@2x.png

The button passes the job to the hypervisor (for VPS-class) or to PXE (for dedicated). The page swaps to a progress panel.

What changes from here

  • A top-of-page banner appears: Your VPS is installing... with a View progress button.
  • The progress panel shows the current step - Preparing, Downloading image, Cloning, Booting, Cloud-init - and a percentage. The portal polls the hypervisor every 2 seconds for fresh status.
  • When the install finishes, the banner clears and the detail page reloads with the new OS shown next to the status pill. The first boot may take an extra minute while cloud-init applies your credentials.
  • The Activity tab records who triggered the reinstall, the chosen image, and the timestamp.

If something looks stuck for more than a few minutes (especially on dedicated), check the location agent status under Management, then contact Support if it doesn't recover.