ebsnvme-id output not consistent

[root]# rpm -qf /sbin/ebsnvme-id ec2-utils-1.0-2.amzn2.noarch [root]# ebsnvme-id -b /dev/nvme0n1 xvda [root]# ebsnvme-id -b /dev/nvme1n1 /dev/xvde [root]# ebsnvme-id -b /dev/nvme2n1 /dev/xvdf We have discovered that if the disks are attached after the instance is created, the command will return the full path. If the disks are attached and then the instance is rebooted for example, the command will prune off the /dev/ prefix. This seems to be a problem with whether or not the ebs volumes are present in the instance metadata. I rebooted one of the instances we were working with yesterday and now the output is reflecting this new output: [root@vlt-us-west-2-production-1106-solr-1 ~]# uptime 13:02:44 up 2 min, 1 user, load average: 0.33, 0.28, 0.11 [root@vlt-us-west-2-production-1106-solr-1 ~]# /sbin/ebsnvme-id -b /dev/nvme2n1 xvde instance doesn't know abt the ebs volume, it was added later [root@vlt-eu-central-1-production-2047-solr-6 ~]# curl -m3 http://169.254.169.254/latest/meta-data/block-device-mapping/ami [root@vlt-us-west-2-production-1106-solr-1 ~]# curl -m3 http://169.254.169.254/latest/meta-data/block-device-mapping/ http://169.254.169.254/latest/meta-data/block-device-mapping ami ebs14 ebs2