On Fri, Jan 5, 2018 at 9:59 AM, Ryan Goulding <ryandgould...@gmail.com> wrote:
> Greetings, James. > > +controller-dev > > Forming a reply has been on my to-do list, but unfortunately have been > quite busy since the holidays. Thanks for reaching out; I actually purely > use gmail now since I have actually now worked on ODL for three different > companies, and having the same email ensures some continuity of > communication with peers. > > I am adding in Tom Pantelis. From private conversations I had with him, > we expose an RPC to do scale out of the cluster. Basically, we never > productized or recommended it commercially, and there is likely some more > scripting needed to update the new "seed" nodes to be in good shape to join > the cluster. I may be butchering terminology here, so I'll defer to Tom > who is the clustering guru of ODL. > > Tom, this is James. James works for Amdocs and was interested in > scale-out of ODL. I have added in controller-dev, as I am sure others > would benefit from hearing how this is done! > > There is an RPC to add a shard replicas and join an existing cluster (implements Raft AddServer RPC in the backend). However it's mainly a building block - there needs to be additional changes to configure the akka.conf et al to talk to the existing cluster etc. Ideally this would be scripted and productized. That said, it isn't really needed for HA, where the cluster members are typically setup once and remain static, which so far has been the use case for ODL clustering, in my experience anyway. Implementing scale-out is a different story - accomplishing that likely depends on the specific use case, application(s) and data models involved with specific sharding configurations on deployment. > Thanks and Best Regards, > > Ryan Goulding > > On Fri, Jan 5, 2018 at 9:42 AM, James MacNider <james.macni...@amdocs.com> > wrote: > >> Hi Ryan, >> >> >> >> I hope that it’s alright that I’m contacting you at your Gmail address, >> it’s the only one I could find on the ONAP wiki. I’d like to follow up on >> the brief exchange we had at the ONAP meeting in Santa Clara where we >> talked about the possibility of scaling an ODL cluster without >> reconfiguring all the nodes and restarting them. I’ve been unable to find >> documentation that reflects this yet, but if I recall correctly, you knew >> the ODL contributor(s) that worked on this feature set. Could you help me >> connect with them to learn more about this? >> >> >> >> Thanks, >> >> >> >> *James MacNider* >> >> Software Architect >> >> >> >> Open Network Division >> >> Amdocs Technology >> >> (office) (613)-595-5213 <(613)%20595-5213> >> >> [image: amdocs-a] >> >> *Amdocs* is a Platinuim Member of ONAP <https://onap.org/> >> >> >> This message and the information contained herein is proprietary and >> confidential and subject to the Amdocs policy statement, >> you may review at https://www.amdocs.com/about/email-disclaimer >> > >
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev
