CentOS 7 Is Dead: Our Real-World Migration to Rocky Linux

CentOS 7 is end of life. No more security patches, no more bug fixes — which means every CentOS 7 box still in production is a growing liability. Here is how we migrate clients off it safely, using Rocky Linux as the drop-in successor.

Why Rocky Linux (or AlmaLinux)

Both Rocky Linux and AlmaLinux are bug-for-bug compatible rebuilds of Red Hat Enterprise Linux, which is exactly what CentOS used to be. For most workloads they are a true drop-in replacement, so your applications behave identically.

In-place upgrade vs fresh build

There are two routes, and we choose based on the server’s age and how much cruft it has accumulated:

• Fresh build + migrate — a clean Rocky install on a new server, applications and data migrated across. Our default: you end up with a documented, known-good box.
• In-place conversion — tools can convert CentOS 7 to a newer release in place. Faster, but riskier on long-lived servers, so we only use it when a rebuild isn’t practical.

The migration playbook

1. Inventory everything: packages, services, cron jobs, firewall rules, users and SSL certificates.
2. Stand up the new Rocky server and replicate the stack, pinning the same major versions of PHP, MySQL/MariaDB and the web server.
3. Sync data with rsync, then do a final delta sync during the cutover window.
4. Test thoroughly on the new box using a hosts-file override before touching DNS.
5. Cut over DNS with a low TTL, monitor closely, and keep the old server ready for rollback.

Minimising downtime

By rehearsing the cutover on a staging copy and lowering DNS TTLs in advance, we typically keep real downtime to a few minutes — often zero for stateless sites behind a load balancer.

If you are still running CentOS 7 (or 8, which is also EOL), the clock is ticking on your security posture. The good news: a planned migration is a calm, boring, well-rehearsed event rather than an emergency.

---

#almalinux #centos 7 eol #os migration #rocky linux