Bug#865965: InvalidBlockDeviceMapping error while creating a new ami based on stretch one
On Mon, Jun 26, 2017 at 09:55:31AM +0200, Xavier Lembo wrote:
> I've tried today to use my jessie working script on the eu-west-1 stretch ami (ami-e79f8781)
> My custom ami must have a specific size for first partition, so , i use a specific block device mapping.
> On the Stretch AMI, it fails with message:
> amazon-ebs: Error launching source instance: InvalidBlockDeviceMapping: The device 'xvda' is used in more than one block-device mapping
> ==> amazon-ebs: status code: 400, request id: ec5463c3-2498-4084-a1aa-825b24b07287
> I've tried to check ebs differences between the working jessie ami and this one
> and the difference is in the naming of the block device:
Are you able to launch a standalone instance of the stretch AMI? I don't
think the issue is with the AMI's block-device mapping specifically.
What's the block-device mapping configuration you're trying to specifiy
with packer? My guess is that you're ending up specifying both "xvda"
and "/dev/xvda" with packer, which is causing the conflict because
they reference the same canonical device.
I've verified that the strech AMI in eu-west-1 does launch correctly
with its current block-device mapping settings:
curl -s 169.254.169.254/latest/meta-data/placement/availability-zone ; echo
admin@ip-172-31-34-141:~$ curl -s 169.254.169.254/latest/meta-data/ami-id ; echo
admin@ip-172-31-34-141:~$ curl -s 169.254.169.254/latest/meta-data/block-device-mapping/ami/ ; echo
admin@ip-172-31-34-141:~$ curl -s 169.254.169.254/latest/meta-data/block-device-mapping/root/ ; echo
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
xvda 202:0 0 8G 0 disk
└─xvda1 202:1 0 8G 0 part /