AWS Application Migration Service (MGN) Windows Servers Experiences

Hey everyone,

Performing a migration of a dozen or so of on premise Windows Servers varying from 2012 R2 to 2019 (mostly 2019). I was curious about other users experiences when utilizing AWS MGN for the lift and shift. 85% of the servers appear to have replicated and booted as an EC2 in AWS during testing without any real issues. Taking these steps: Installing the MGN agent, performing an initial replication, launching tests, confirming the EC2 Instance boots to login screen (instance screenshot and rdp). There’s about 15% of the servers, I’ve had to work with AWS support to figure how why they aren’t launching. They have recommended using an IAM policy to force conversion servers onto a non-nitro instance, switching boot mode from Legacy Bios to UEFI (if applicable) in the Source Server launch settings, trying different instances types, etc. After working with support to troubleshoot, we can get the instances to a working state for testing. A couple weeks ago by (source servers still replicating) and we’ll do another round of testing. The 85% work as they normally do and the 15% we needed to find tune (changes from prior troubleshooting sessions applied during this round of testing) will have new problems and we’ll open support tickets again as they will not boot up or even blue screen. I’ve been advised these servers are all working fine on premise, but I can’t see that for myself. On cutover day we’ll go into it with a toolbelt of troubleshooting steps, but my confidence in this AWS service is dwindling by the day.

I was wondering if others have experienced anything similar using MGN to lift and shift.

Thanks!