Hello Noah
the problem does not come from Packer, actually AWS allows one to create images where the root device is either xvda or /dev/xvda. This does not pose any problem for booting as you did, but it does triggers this error message from AWS's API even if there is no actual conflicts. The message is misleading. I got the exact issue when creating spot fleet requests wich awscli. Replacing /dev/xvda with xvda in the spot fleet configuration fixed the issue. More reading about that issue : https://github.com/capitalone/cloud-custodian/issues/1524#issuecomment-326539507 So the final word is that Debian should be careful when creating images because the device naming does matter. It can be either xvda or /dev/xvda, but changing it also requires users to update their configuration to prevent AWS's API from being confused. Regards, dud
