[Pacemaker] Resource cannot run anywhere

[Lars Ellenberg]

On Mon, Sep 15, 2008 at 10:03:17AM +0200, Lars Marowsky-Bree wrote:
> On 2008-09-13T22:57:03, Lars Ellenberg <lars.ellenberg at linbit.com> wrote:
> 
> > > There were _reasons_ why I insisted that drbd8 doesn't work with
> > > pacemarker yet. ;-)
> > 
> > why not?
> > it does work fine.
> 
> Not really. The RA doesn't know about "outdated" for example and would
> still flag that node as eligible for promotion.

who cares.
it would refuse.
looks like "failed promotion" then.
what "recovery actions" would that cause?
what harm would that do?

> > you just have to skip the "floating peer" nonsense^Wstuff.
> 
> No; in fact, I think you have to skip the m/s approach and use drbddisk
> instead, and also forego monitoring of the secondary.

that is still pacemaker. means, it works with pacemaker.

besides,
as currently implemented in the RA,
what exactly is that monitoring good for,
I mean what does it protect me against?

what _real_ (not theoretical, somewhen, iff implemented by someone)
benefits would I have (don't mention floating peers) from choosing the
current drbd OCF RA over drbddisk _now_?

-- 
: Lars Ellenberg                
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.