1. Messages
  2. Running NATS in K8s

Re: Running NATS in K8s

Dan Kohn <dan@...>
 

Regarding cross-cloud, the project is here: https://github.com/cncf/cross-cloud/ and I will make intros as NATS goes forward.

Echoing your previous points, a lot of what the cross-cloud team spends time on is the continuous deployment packaging of projects into a chart, so if NATS moves forward on that work it will dramatically simplify what cross-cloud needs to do.

--
Dan Kohn <dan@...>
Executive Director, Cloud Native Computing Foundation https://www.cncf.io
+1-415-233-1000 https://www.dankohn.com

On Tue, Oct 3, 2017 at 1:41 PM, Bob Wise via cncf-toc <cncf-toc@...> wrote:
Peter,

Thank you.

I would advocate for building a chart for deployment on Kubernetes in the short run and an operator (much bigger job) over time.
I would advise that your team should consider the chart as first-class maintained artifact as an important step for the core NATS server as well.

My team (Samsung CNCT) have done a lot of work around charts, and happy to offer some advice/assistance.
I can connect you with our charts lead (Matt Farina) who is also one of the Kubernetes SIG Apps leads.

Noting that the pires project referenced below appears to be tested against 1.5.x. That's getting pretty old in Kubernetes releases.
If you are interested in nats-on-kubernetes CI (which I think you should be), I can also offer assistance via one of our
other team members - Aaron Crickenberger.

CNCF also has a growing cross project CI effort that we would recommend you get involved in. Chris - do we have a normal first point of contact for those interested in this?

Offer is open to anyone on the candidate project lists needing some opinionated guidance.

-Bob Wise


On Tue, Oct 3, 2017 at 9:29 AM, Peter Miron via cncf-toc <cncf-toc@...> wrote:
As requested on today's call.

Core NATS server on Kubernetes (with custom route generator):
https://github.com/pires/kubernetes-nats-cluster

NATS Streaming with NFS (Helm Chart):

Also, a quick note on NATS Streaming, we do have file-locked based fault tolerance today, as configured in the above example. This relies on you having trusted, replicated storage vs. the NATS server handling redundancy.

If anyone has suggestions for how to more easily configure with some of the more recent container storage specs and possibly replica sets, let us know. Happy to hear ideas. Thanks!
Pete

_______________________________________________
cncf-toc mailing list
cncf-toc@...
https://lists.cncf.io/mailman/listinfo/cncf-toc



_______________________________________________
cncf-toc mailing list
cncf-toc@...
https://lists.cncf.io/mailman/listinfo/cncf-toc
Join cncf-toc@lists.cncf.io to automatically receive all group messages.