CroZone
Release wireHermes Agent v0.23.3  ·  Stable channel patchImmich v3.4.2  ·  Library monitoring fixes
Guides

Bring up an Ubuntu VPS and prove you can recover it

A first-hour walkthrough for a disposable server, from SSH login to a working page and a recovery check.

CroZone · Published Nov 20, 2026
In this article

Provision a small test instance

Choose an Ubuntu LTS image, a region and a plan with a known hourly or monthly charge. Add your public SSH key during provisioning. Restrict inbound SSH to your intended administration address in the provider firewall. Locate the web console and read how snapshots are billed before creating one.

Verify the host and the update path

Use the provider console to check the host key fingerprint before accepting it on your first SSH connection. Use the username supplied by the image. Inspect disk space, failed services and the automatic-update timers. A timer listing shows scheduling; inspect the unattended-upgrades logs to confirm what actually ran.

df -h /
systemctl --failed
systemctl list-timers --all "apt-*"

Serve a page over an SSH tunnel

In a fresh test folder on the VPS, start a loopback HTTP server. In a second terminal on your computer, replace USER and HOST and open the tunnel. Visit http://127.0.0.1:8081 locally. Your SSH session carries the page without a public HTTP firewall rule.

mkdir -p ~/vps-demo && cd ~/vps-demo
printf '%s\n' 'VPS recovery check' > index.html
python3 -m http.server 8080 --bind 127.0.0.1

# In a terminal on your own computer:
ssh -N -L 8081:127.0.0.1:8080 USER@HOST

Exercise recovery and end the trial

Stop the test server with Ctrl-C. Take a provider snapshot and restore it to a separate temporary instance. Check that you can sign in and read the test page from disk. Delete disposable instances, disks and snapshots when finished, checking the provider billing page for retained resources.

Decide what can fail independently

Write down the files you created and who supplies them. Ubuntu and system packages can come from a fresh image. A deployment manifest can come from version control. App data needs its own backup. SSH credentials and data-encryption keys need a recovery location that survives the loss of the server. On a single VPS, it is easy to put every copy on the same disk without realizing it.

The tunnel test answers a narrower question than a public deployment. It proves the machine can run the app and that an authorized operator can reach it. If it fails, first run curl from the VPS against its local address. If that works, inspect the local SSH session for an already-used port or forwarding policy. Keep one working SSH session open during routine access configuration and prove a second login before closing it.

Make the snapshot drill falsifiable

Before restoring the snapshot to another instance, put a date and a random test string in the page. After boot, the restored instance should show the same text when read from disk. Check which firewall rules and disks followed the restore and which your provider expects you to recreate. Track time from starting the restore to reading the file. For a real service, add an app-level verification step and test restoration of its independent backup. A disk snapshot is a point-in-time copy. The restore drill tells you whether your actual service can use what was captured.

Primary sources

Ubuntu Server · Automatic updates

Read next   Local or cloud? Follow the work.

Search CroZone