OS reload failed with "Invalid template" — what it means and how to recover

You ran an OS reload from the customer portal and the result came back: Service: NNNN, Distro: <name>, Error: Invalid template. This article explains what that message actually means, the most common causes, and how to get unstuck.

What "Invalid template" means

OS reload pulls a base disk image (a "template") for the distro you selected, then writes it onto your VPS's storage. "Invalid template" means the template the portal tried to deploy isn't available or isn't compatible with your VPS's configuration. Three flavours:

• The distro is retired. Old distros (Debian 8, Ubuntu 14.04, CentOS 7, Slackware 14.2) are eventually pulled from the active template list. The portal sometimes still shows them in the dropdown for backward compatibility, but the underlying image is gone.
• The template doesn't fit your disk size. Some templates require a minimum disk allocation. If your VPS has a smaller disk than the template's minimum, the deploy fails.
• The template's architecture doesn't match. Mostly historical — 32-bit templates on a 64-bit VM, or vice versa. Almost never seen on modern orders.

Pick a different distro

The quickest fix: try a different distro from the dropdown. Currently supported choices for Linux VPS:

• Debian 12 (current stable) — most flexible, good package coverage.
• Debian 11 (oldstable) — still supported; use if you have legacy software pinning.
• Ubuntu 24.04 LTS — best for newer kernel features and faster moving packages.
• Ubuntu 22.04 LTS — still supported.
• AlmaLinux 9, Rocky Linux 9 — for RPM-based workloads.

If you specifically need an older distro for software compatibility, see "If you need an older distro" below.

If you need an older distro

Sometimes you legitimately need an older release — vendor software that hasn't been updated, an old PBX distro your customer requires. Options:

• Open a ticket. Some retired templates are still in the archive and can be deployed by a human on request. Specify exactly which distro and version, and what you need it for (helps us check viability).
• Install on a current distro then upgrade in-place. For Debian/Ubuntu, install a current version and downgrade specific packages. More work but avoids running an unsupported base.
• Use a container. Run the legacy software in Docker or Podman on a modern host. The host distro is current, the software runs in its own environment. See Docker on a small VPS.

If the disk is too small

Some distros expect a minimum disk — typically 20 GB. If your VPS has less, the template won't deploy. Options:

• Pick a smaller distro. Plain Debian or Alpine fits in tiny footprints; full-featured distros with desktop environments or PBX stacks don't.
• Upgrade disk allocation first, then retry the reload. See Upgrading your VPS plan.

If the same distro worked before

If you've successfully reloaded to (say) Debian 11 on this VPS before and now it fails, the template was retired or moved. Open a ticket — we can either re-stage the template for you or recommend a successor.

What the OS reload actually does

Worth knowing what you're about to commit to:

• Everything on the VPS is destroyed. Filesystem, configs, data — gone. The new template is a clean OS install.
• The IP, hostname, and account stay the same. Same SSH endpoint, same DNS records continue to work.
• You get a new root password in the welcome email.
• Any snapshots you took stay around for a while, depending on storage. If you didn't take a final snapshot before reloading and you want your old data back, open a ticket immediately — the sooner the better.

Take an off-host backup before the reload. The portal's snapshot feature stores on the same storage as the VM; an OS reload may or may not preserve those snapshots depending on the hypervisor specifics. See Off-host backup strategies.

If the reload says "complete" but the VPS won't boot

Different failure mode: the reload reports success, you log in to the console, and the VM doesn't boot or boots to a broken state. Causes:

• Template is broken upstream. The image itself is corrupted or missing a kernel module needed for the hypervisor. Open a ticket; we'll swap it.
• Disk wasn't fully wiped before the new image landed. Rare, but possible if the previous OS had unusual partition layout. Boot to rescue mode (see Rescue mode) and re-partition manually, or just request a second reload.