Linux server migration planning

If you run some number of Linux servers for yourself you'll inevitably need to migrate one to a new machine at some point. This might be straightforward, particularly if you can do so by migrating a VM image or if you treat your personal servers like cattle, not pets. If, however, you use your servers like I do, the migration process is not trivial and takes some planning.

This post attempts to index everywhere I check to see what a server is running, plus other stuff to consider when completing the migration. The result can be used to make a migration plan.

What is the server running?

• systemd services
• services running under Docker
• check tailscale serve status
• Let's Encrypt certificates (/etc/letsencrypt/live)
• Nginx websites (/etc/nginx/sites-enabled)
• check /etc/cron* and user-level crontabs
• anything interesting in /home
• anything interesting in /var/lib (such as Docker, databases, etc)
• anything in /srv
• check netstat -lnptu
• check ufw status
• what targets hosted on this machine are your monitoring systems currently monitoring?
• are there other places that you use that aren't covered here? (for example, PM2, or software installed in /opt?)

I've written a script that indexes most of the stuff listed above. If nothing else, it's a helpful starting point. You can run it directly from the linked Gist via:

curl -s https://gist.githubusercontent.com/cdzombak/dfa620c01bc69f7f419906c3b5c62800/raw/3d3cb09f9935377d22e06964df6243967c4fe503/sysindex.sh | sudo bash

Other stuff to remember

• DNS - update for the new server; cleanup old entries
• Reverse DNS - set for the new server
• Monitoring - monitor the new server & services as you stand them up (I use Uptime Kuma for this)
• keep your password manager up to date & tidied as you perform the migration