… admittedly I can’t remember if I followed those particular instructions, but I did successfully deploy on the lowest-end OVH VPS (value-1-2-40) about a year ago.
basically:
- order a VPS, I went for debian as that’s my usual, wait for it to be provisioned
- if you care for ipv6, note it down because ovh’s issuing them as static config to the predeployed linux via the cloud-init tooling (ipv4 is dhcp)
- boot to rescue mode, open the kvm web console
- /dev/sda should be the rescueOS-disk, /dev/sdb should be your VPS persistent disk, check with cfdisk for example
- just in case, zeroize: dd if=/dev/zero of=/dev/sdb
- wget the CHR image to /tmp, unpack, dd over
- just in case, sync
- exit rescue mode using the control panel (don’t just reboot the rescue OS from console)
- CHR should boot, automatically resize to take advantage of the full size of your VPS disk, then reboot again
- turbo log in and set a password before the first botnet scanner finds telnet + ssh naked open and moves in
I successfully got one running on Linode as well, the trick over there is that the default boot config assumes disk images are raw filesystems while kernel is fed from host.
So you need “Select Kernel: Direct Disk” and disable all of the “boot helpers” at far bottom of the VPS bootconfig page.
On the upside, you can make the first boot with ethernet “unplugged” to not worry about racing the bots for first login.