EC2 Ubuntu Instance Unreachable - Instance Reachability Check Failed for AMI-based Instance"

When it boots up (as much as it can) is there anything on the system console that shows how far it has got? e.g. does it look like it's sitting in maintenance mode, or at a GRUB menu, or has it panicked and there's a panic string, or anything else? Select the instance and choose Actions -> Monitor and troubleshoot -> Get instance screenshot

When working with the original instance (the one you created your custom AMI from) did you make any major changes to it before creating the new AMI? I'm talking so much about installing the odd bit of software (although that can't be ruled out as a cause), but more like did you grow the root disk, or put /var on a separate slice, or bring things under LVM control where previously they weren't, or change SELinux from permissive to enforcing, that kind of thing.

Also check there aren't situations like a full / or /var filesystem on the original instance - while that one will carry on running and you may not notice anything, a new AMI created from a system in that state will have problems from the off.

Next step would be to stop the problem instance and detach its root volume. Then go back to the original instance (if it's still running, if not then provision a new instance from scratch with the plain AMI) and once it's booted attach the root disk from the problem instance as /dev/sdf (I think that's it, whatever the default value is).

Then on the command line run lsblk to identify the newly-attached device, and mount it on /mnt (if /var or /var/log is a separate filesystem then mount that one instead). This allows you to go through the problem instance's logs in /mnt/var/log. You may even want to tar or gzip these up and keep them somewhere that you can go though them later (like an S3 bucket). For a quick win it would also be worth sanity-checking /mnt/etc/fstab.

Also take a look at https://repost.aws/knowledge-center/ec2-linux-emergency-mode for more suggestions.

Is the IP address of your source instance hardcoded in the configuration file? If your IP address is hardcoded in a configuration file, you can encounter issues that cause your network to fail to come up. This occurs when an Amazon Machine Image (AMI) is taken from an instance with a statically configured IP address. To fix this error, set your network interface on your source instance to use DHCP and create a new AMI.

Note that you can't update existing AMIs. Setting the network interface to use DHCP has to be done on the instance before creating a new AMI.

Hi Ghazi,

Take a look at this article: https://repost.aws/knowledge-center/ec2-linux-status-check-failure.

Instance status checks may fail for the following reasons. You can use the following methods to troubleshoot an unreachable Linux instance.