[Pacemaker] migrate_from/migrate_to for Stateful RAs

Andrew Beekhof andrew at beekhof.net
Tue Sep 1 04:16:31 EDT 2009


On Tue, Sep 1, 2009 at 8:50 AM, Florian Haas<florian.haas at linbit.com> wrote:
> On 09/01/2009 08:30 AM, Andrew Beekhof wrote:
>> only one layer of the stack is ever able to migrate.
>>
>> a resource can only migrate if its location and ordering pre-requisits
>> are satisfied on the source and target nodes.
>> clearly this can't be true for drbd _and_ something running on top of
>> it (since drbd wouldn't be on the target yet).
>
> That's what I was afraid of. :) This leaves me with handling DRBD state
> transitions at the resource level like we currently do for Xen
> (http://www.drbd.org/users-guide/s-xen-configure-domu.html) which is not
> generic, has a lot of caveats (like support from pygrub and qemu-dm),
> and most importantly, forces me to take control over DRBD away from
> Pacemaker.
>
> I had hoped there would be a better solution. Seems like there isn't. :(

"yet"




More information about the Pacemaker mailing list