In EC2 there are two ways to "copy" a machine and it is important not to confuse them. An EBS snapshot is a copy of one disk. An AMI (Amazon Machine Image) is a copy of the whole machine: all its disks plus the information needed to boot. With a snapshot you recover data; with an AMI you launch an identical new machine in three clicks. If you want a backup "you can boot from", what you are after is an AMI.
This guide creates the AMI, launches a copy from it and, what almost nobody does, cleans up afterwards.
Before you start: what each thing means
- EC2
- Amazon's "made-to-measure" servers, paid for by the hours they are running.
- Instance
- Each EC2 machine or server.
- AMI
- Amazon Machine Image: a complete copy of the machine (all its disks plus what is needed to boot) from which identical new machines can be created.
- EBS snapshot
- The copy of a single disk. An AMI is made up of one or more EBS snapshots, which are what takes up space and gets charged.
- Launch
- What AWS calls creating and starting a new machine.
- Security group
- The AWS firewall for a machine: from where and through which ports you can connect. The copy doesn't inherit it; you have to choose it.
- SSH key pair
- The key a technician uses to log into the machine from a terminal. When launching the copy you choose which one to use.
- Elastic IP
- A fixed internet address that can be moved from the old machine to the new one so nobody notices the change.
- Data Lifecycle Manager
- An AWS service that makes backups (AMIs or snapshots) on a schedule and deletes old ones on its own.
Step by step
Decide whether to stop the machine or not
EC2 can create the image with the machine running. By default it reboots the instance briefly to make sure the disk is consistent; you can tick No reboot to avoid the interruption, accepting the risk that some open file ends up half-written. For a website with a database, pick a low-traffic moment and let it reboot, or stop the database for a minute and use "no reboot".
Create the image
In the EC2 console, Instances, select the machine, Actions, Image and templates, Create image. Give it a name with the date and reason (
web-tienda-antes-migracion-2026-10-26) and a description. Review the list of volumes: all the machine's disks appear and you can exclude some (a temporary data disk, for example). Tick the Delete on termination box on the ones you want to disappear when a machine launched from this image is terminated.Wait for it to become available
In the AMIs menu (under Images) you will see it as "pending" and then "available". How long it takes depends on the size of the disks. Notice that one or more new EBS snapshots have appeared under Snapshots: the AMI relies on them and they are what actually takes up space and gets charged.
Launch a copy from the AMI
Select the AMI, Launch instance from AMI. Choose the instance type (it can differ from the original: here you can grow or shrink in CPU and memory), the SSH key pair, the network and, very important, the security group: the copy doesn't inherit the original machine's, so choose the same one. Launch. In a couple of minutes you have a twin machine.
Check it and decide what to do with it
Log in via SSH and check that the services start. Typical uses: a test machine identical to production, replacing a broken machine, or migrating to another instance type. If the copy is going to replace the original, move the Elastic IP from one to the other and, when you are sure, terminate the old one.
Clean up old images
AMIs don't expire and each one drags its snapshots along. Under AMIs, select the ones that no longer make sense and Deregister AMI; then go to Snapshots and delete the ones that belonged to that image (the description says so). If you want this to happen on its own, Amazon Data Lifecycle Manager creates images every night and deletes those older than N days without you having to remember.
What can go wrong
The copy doesn't boot or doesn't respond
Look at the instance's screenshot (Actions, Monitor and troubleshoot) and the system log. The most common causes: the image was made "no reboot" at a bad moment, or the security group doesn't open the port you are trying to get in through.
I deleted a snapshot and the AMI no longer works
The AMI depends on its snapshots; without them it can't launch anything. Deregister that AMI too so it doesn't confuse you. The lesson: deregister first, delete snapshots afterwards.
The copies have the same hostname and the same keys
Normal: they are clones. If you are going to have the original and the copy running at the same time, change the hostname on the copy and, if the machine carries agents or licences tied to the identifier, reinitialise them.
I want the copy in another region
Select the AMI, Actions, Copy AMI, and choose the destination region. When it is available there, you launch from that copy. Useful as a contingency plan against a regional problem.
In Simplio
In Simplio EC2 machines appear alongside Lightsail ones with their metrics, their snapshots and the start, stop and reboot buttons; backups are scheduled per machine (every day, every week, how many to keep) and can be copied to another region automatically, and if a scheduled backup doesn't happen, it warns you. And each person on your team sees and touches only the machines you assign to them.
The same task, in AWS and in Simplio
| In the AWS console | In Simplio | |
|---|---|---|
| Copy of the machine | EC2 console, Actions, Image and templates, Create image, wait: 5 clicks | Snapshot button on the machine: 1 click |
| Periodic backups and cleanup | Set up Data Lifecycle Manager with policies | Per-machine schedule with how many to keep; old ones are deleted on their own |
| Copy in another region | Copy the AMI by hand every time | "Copy to another region" box in the schedule |