Open Stack

On 5/8/2018 7:41 AM, Bob Ball wrote:
> I'd be hesitant to permit reboot-from-rescue for all drivers as I'm not sure the drivers would have consistent (or perhaps working!) behaviours?  Is there a way to enable this when using XenAPI?

Off the top of my head the virt driver could report a capability for 
this which gets modeled in placement as a standard trait on the compute 
node resource provider. Then the API could check for that trait from 
Placement and fail if it's not found.

-- 

Thanks,

Matt