On aws with nvme the script always adds after reboot wrong disk on lvm

[cristifalcas]

On m5 instances, with nvme:

• configure docker-storage-setup with the real disk name (nvme1n1 in this case)
• run container-storage-setup. Everything is fine for now
• add more ebs voluyme to instance
• reboot instance
• now another disk is nvme1n1, and the original one is, for example, nvme5n1
• container-storage-setup adds the new disk also to the lvm, by wiping the disk completely. Of course, all data is lost now on that disk

# cat /etc/sysconfig/docker-storage-setup
DEVS=nvme1n1
GROWPART=true
ROOT_SIZE=+100%FREE
STORAGE_DRIVER=overlay2
VG=docker
CONTAINER_ROOT_LV_NAME=pool00
CONTAINER_ROOT_LV_SIZE=100%FREE
CONTAINER_ROOT_LV_MOUNT_PATH=/var/lib/docker
POOL_META_SIZE=1G
WIPE_SIGNATURES=true
EXTRA_STORAGE_OPTIONS="--storage-opt overlay2.size=20G"

Is there any way to resolve this?

[npoyant]

problem still exists