Amazon Linux 2 reached end of support on June 30, 2026. Servers still running it keep working, but they no longer receive security updates, which is the kind of finding that fails audits and insurance questionnaires. The replacement is Amazon Linux 2023. There is no in-place upgrade between the two: moving means building new servers and moving the workload across, which is also the cleanest way to do it.
Find servers still on Amazon Linux 2
- On a server, run
cat /etc/os-release. Amazon Linux 2 showsVERSION="2"; Amazon Linux 2023 showsVERSION="2023". - In the EC2 console, check each instance's AMI name; Amazon Linux 2 images are named
amzn2-ami-.... - AWS Systems Manager inventory, if enabled, lists operating system versions across all servers at once.
- Do not forget build servers, bastions, and old test instances.
What is different in Amazon Linux 2023
| Area | Amazon Linux 2 | Amazon Linux 2023 |
|---|---|---|
| Package manager | yum and amazon-linux-extras | dnf; no extras catalog |
| Updates | Rolling | Versioned repositories, so a server stays on a known package set until you choose to move |
| SELinux | Disabled by default | Enabled, in permissive mode by default |
| Instance metadata | IMDSv1 and v2 | IMDSv2 by default |
| Software versions | Older PHP, Python, and other runtimes through extras | Newer runtimes from the main repositories |
Check that every package your site needs, such as a specific PHP version, image library, or monitoring agent, is available on Amazon Linux 2023 before starting.
Rebuild, don't upgrade
- Write down the current server: installed packages, configuration files, scheduled jobs, users, and firewall rules. Better still, check whether the build is already captured in Infrastructure as Code.
- Build new servers on Amazon Linux 2023 from code, with the same configuration adapted to the new system.
- Test on a temporary address: site rendering, sign-in, forms, scheduled jobs, email sending, and integrations.
- Switch traffic at the load balancer, keeping the old servers stopped but available as the rollback.
- Remove the old servers once the new ones have run cleanly.
Things that commonly need attention
- Software from extras: anything installed with
amazon-linux-extrasmust come from somewhere else on Amazon Linux 2023, usually the main repositories. - Cron and scheduled tasks: recreate them, and confirm the time zone and paths.
- SELinux in permissive mode: check the audit log for denials before ever switching to enforcing.
- Scripts that call the instance metadata service without a session token stop working under IMDSv2 only.
- Monitoring and backup agents must be installed and registered on the new servers.
- Allowlists elsewhere: if partners allowlist your server's address, keep the same address or tell them in advance.
Common questions
Can we pay for more support for Amazon Linux 2? Plan on moving instead. Unsupported operating systems are a recurring audit finding, and each month adds risk.
Should we move to a different distribution instead? That can make sense if you already standardize on Red Hat Enterprise Linux or Ubuntu; see moving off end-of-life Linux distributions.
Do containers need this too? Container images built on Amazon Linux 2 base images need rebuilding on a supported base.
How long does a rebuild take? If the server is already defined in Infrastructure as Code, mostly testing time. If it was built by hand years ago, documenting what is on it is the biggest part of the work.
Will the site go down? It should not. New servers are tested on a temporary address and traffic is switched at the load balancer, with the old servers kept as the rollback.
Can we run Amazon Linux 2 a little longer while we prepare? Every month on an unsupported system means security fixes you are not getting; plan the move as soon as possible and keep the remaining servers tightly restricted until then.
Does the new system cost more? No: the operating system itself has no extra charge, and new servers are often a chance to right-size.
How eWay handles operating system moves
On environments eWay manages, operating system versions are moved in step with what the application supports, and servers are built from hardened images through Infrastructure as Code, so a rebuild on a new operating system is a planned change rather than a hand-crafted server. Changes are validated in a test environment first, patching happens one node at a time behind a load balancer, and every change has a tested rollback.
What to include in a support ticket
- The servers or environments still on Amazon Linux 2
- Any software with specific version needs
- Dates to avoid for the switch
eWay customers can open a ticket in the support portal or email support@ewaycorp.com. Related reading: keeping servers patched and SELinux.
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article