[Linux-HA] Question regarding the stonith resources

[Andrew Beekhof]

On 10/31/05, Alan Robertson <alanr at unix.sh> wrote:
> Peter Kruse wrote:
> > Andrew Beekhof wrote:
> >> On 10/31/05, Peter Kruse <pk at q-leap.com> wrote:
> >>
> >> really?
> >> surely there is a way to say (for each stonith resource) "i can shoot
> >> node X, Y and Z" no?
> >
> > The parameters for the apcmastersnmp I found are:
> > ipaddr, port and community.  I guess the plugin
> > can find out the outlet names, which are the nodenames.
> > But it is not clear what happens if it cannot
> > find an outlet name it expected to find...
> >
> >>> right, can it happen that the DC shoots itself but doesn't succeed
> >>> and the other node still takes over?
> >>
> >> no - if its still around - then its still the DC
> >
> > Ah!  so it's implicit:  if it succeeds, then it's gone
> > and the other node can take over, if it doesn't then
> > you get an error.  But that also means per default
> > success is assumed.  "gone" in the above statement
> > is defined as "loss of communications", right?
> > Ok, that sounds reasonable as the communications
> > are redundant and losing all communications is
> > at least a double failure which is not covered...
> > Guess I can live with that if I haven't overseen anything...
>
>
> It's simpler than that.  If a node dies, all resources it owns are
> considered to have been stopped.
>
> If it doesn't die (according to heartbeat), then you can'trestart its
> resources, because they're still running.
>

exactly :-)