I’ve spent the better part of a week working on this, and we have finally found all the little issues, so far…

Here’s the problem, we had a citrix server in AWS running some services for monitoring for our company, which was overloading the server causing usability issues both with citrix and the services. Rather than building a new server from scratch and reconfiguring either citrix or the services, we figured that we would just spin up a new instance from a nightly backup of the original server and remove components from each, which would be quicker than starting from scratch, theoretically that is. After way too much time spent on finding each glitch, here is the end result.

1. Spin up a new instance that will be your duplicate. NOTE: if you are in a domain, to avoid conflicts, put this server in its own security group so it cannot see the domain and cannot conflict with your live server.

2. Let it boot all the way up and connect to it. Once logged in use the EC2config service to rename the system on boot and set the admin password you desire. For more info on this service, see this link. Once those parameters are set, shut the instance down.

3. Take the latest snapshot in the AWS console for the instance you are copying, right click it and create a volume from it.

4. Detach the instance’s volume that it spun up with and attach the volume created from the snapshot on /dev/sda1 to make it the boot drive.

5. let the instance completely boot up. At this point, it will have the same network interface and IP as the original server, which can’t be changed in a VPC.

6. Shut the instance down and attach a new network interface with the desired IP address. Remove the old interface. NOTE: I did not do this portion personally, so I’m not sure if you need to boot it then shut down again to remove the old interface, but I would assume you can’t do it live.

You will now have a duplicate instance. It seems easy now that we have the steps, but getting here was such a pain I would have rather started from scratch.