[Planetlab-devel] migrating OneLab sites to PlanetLab Europe

[Larry Peterson]

Go ahead and move UCL. A brief posting to users is probably appropriate.
I plan to send an announcement (and post something on the web page)
about federation real soon now.

ACK regarding Pisa (we'll let you know as soon as the paperwork clears).

Larry

On 8/13/07, Timur Friedman <timur.friedman at upmc.fr> wrote:
> Dear colleagues at Princeton,
>
> After several weeks of PlanetLab Central - PlanetLab Europe federation
> with just a few PLE sites (INRIA, UPMC, Polish Telecom) and nodes, we're
> ready to start migrating other members of the OneLab project (there are
> ten institutions altogether).
>
> Our first test case is the Université Catholique de Louvain (UCL), which
> is ready to go.
>
> In the database, you should now see two entries for UCL:
>
> - louvain (PLC)
> - uclple (PLE)
>
> If you have no objection, we'll give UCL the go-ahead to take their
> nodes down from PLC and put them up on PLE.
>
> If you think it best, we could send a message alerting users that
> slivers on these nodes will disappear. (In this particular case, it
> looks like just one node from the PLC-based site is active right now.)
> Otherwise, we could just shut down the nodes and bring them back up.
>
> Then, if it's alright with you, they'll keep their current PLC-based
> slices until they're ready to shut them down. We've asked our partners
> to take care not to create more then their allotment of slices, even
> though, during the transition period, they'd technically be able to do so.
>
> Once their experiments are done in the PLC-based slices, we can go ahead
> and close the "louvain" site, and the migration will be complete.
>
> Another test case is the University of Pisa (UniPi), which has entered a
> join request at PLE, with the login base 'unipi'.
>
> If I'm not mistaken, the University of Pisa is not yet a member of
> PlanetLab.  If this is correct, they'll have to become members before we
> approve their join request.  We'll wait to hear from you that it is OK.
> At that point, since there is no question of migration, they can simply
> bring up their nodes as soon as they are ready.
>
> Best regards,
>
> Timur
>
> _______________________________________________
> Devel mailing list
> Devel at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/devel
>
>