[ClusterLabs] Antw: [EXT] Re: failed migration handled the wrong way

[Andrei Borzenkov]

On Mon, Feb 1, 2021 at 1:59 PM Ulrich Windl
<Ulrich.Windl at rz.uni-regensburg.de> wrote:
>
> But the VM *wasn't* stopped on h16!
>

I am not sure what you mean here. It was not stopped during migration?
Yes, pacemaker knew it and it tried to stop it explicitly when
migration failed. It was not stopped when pacemaker tried to stop it?
It is up to resource agent. Resource agent returned success from stop
operation - how should pacemaker know that nothing has actually been
stopped?