You would need to extend Jupyter Notebook server to request and connect to remote kernels. It is very possible and we've done it internally. The component in the Notebook Server that you would need to extend is called the KernelManager.
Another way to remote a Toree kernel is by using https://github.com/jupyter-incubator/kernel_gateway. This project is very close to graduating from incubation. On Fri, Mar 4, 2016 at 4:17 AM, Gregor Moehler <[email protected]> wrote: > Hi, > > the mission statement of toree states that it can be used to submit jobs > on a remote cluster. I see how this can be achieved using the client > library. > > Unfortunately I have not come across examples where spark-kernel/toree is > used to connect Jupyter/Notebook with a spark cluster running remotely. > > Do you have any material that would ease this task? > > Best wishes, > Gregor > > > Gregor Möhler > > > dashDb & Netezza Development > IBM Analytics > Phone: +49-(0)7031-16-5402 > IBM Deutschland Research & Development GmbH > Schönaicher Str. 220, 71032 Böblingen > > Vorsitzende des Aufsichtsrats: Martina Koederitz > Geschäftsführung: Dirk Wittkop > Sitz der Gesellschaft: Böblingen > Registergericht: Amtsgericht Stuttgart, HRB 243294 > > > >
