backend reset should be later in the order of playbooks executed.

[kristvanbesien]

To cleanup my experiments I would like to do that with a small reset config...

Eg.:

[hosts]
datanode1
datanode2
arbiternode

[volume1]
action=delete
volname=volume2

[volume2]
action=delete
volname=volume1

[backend-reset]
pvs=xvdb,xvdc
unmount=yes

This however fails, as the backend-reset is attempted first. gdeploy should first remove the volumes...

[johnsimcall]

I have the same problem. I was expecting the actions to be executed similar to ansible, from top to bottom, but that is not the case. "backend-reset" is always executed first. This causes problems when you're trying to cleanly umount the volume, delete volume, then clean lv/vg/pvs

[sac]

@kristvanbesien @johnsimcall earlier there was a requirement that backend-reset had to be kept separate since it is destructive operation. And had to be run separately. However, now I see that it is more of a hindrance. I will work on getting this integrated.