rely on udev to manage symlinks, closes #2

[brahman81]

• uses xvd notation to match Ubuntu default for Xen virtual devices
• uses [:space:] character set to delete all whitespace
• move script to /usr/local/sbin as this is a system script

[zymergen-vdc]

The name of this file doesn't match the udev rule reference.

[brahman81]

Good catch! Just submitted a fix.

[Kasama]

It would probably be nice to update the README as well to make sure the explanation match the actual provided files.

[b-dean]

Using this PR I still get the wrong devices for the root device:

[root@ip-10-75-0-84 ec2-user]# ls -l /dev/xv*
lrwxrwxrwx 1 root root 7 Mar  9 20:07 /dev/xvda -> nvme1n1
lrwxrwxrwx 1 root root 9 Mar  9 20:07 /dev/xvda1 -> nvme0n1p1
lrwxrwxrwx 1 root root 7 Mar  9 20:07 /dev/xvdb -> nvme2n1

when it should be /dev/nvme0n1p2

[root@ip-10-75-0-84 ec2-user]# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/nvme0n1p2   10G  3.2G  6.9G  32% /
devtmpfs        3.7G     0  3.7G   0% /dev
tmpfs           3.7G     0  3.7G   0% /dev/shm
tmpfs           3.7G   17M  3.7G   1% /run
tmpfs           3.7G     0  3.7G   0% /sys/fs/cgroup
tmpfs           753M     0  753M   0% /run/user/1000

(note that I don't have the xvdb mounted anywhere, was just adding another block device)

The vendor specific info says sda1 for the device that ends up on /. This is a RHEL-7 AMI.

Should there be a /dev/xvda2?

[oogali]

I assume this comment block was from your gist, I don't think this needs to be incorporated upstream.

[oogali]

If your test conditions are not matched, this script will silently fail (with a return value of zero).

A better option would be to invert the test conditions, as to be loud and verbose that we can't find the nvme program and that a required argument is missing.

[oogali]

If the required argument passed is not a valid path to a block device -- should we complain? Or should we let nvme complain and have the user decipher its error message?