On Wednesday, October 25, 2017 at 5:25:37 PM UTC+5:30, Matthias Rampke wrote: > Don't use a loadbalancer for etcd, just tell each kube-apiserver about all 3. > Use DNS names if you don't want to update all apiservers whenever you swap > out an etcd, but that's optional. > > > For the apiserver itself, the simplest way would be to just put all 3 behind > a single DNS name. Otherwise, use any loadbalancer that is available to you, > it doesn't really matter, as long as it accepts long-lived connections. I > think session affinity is not required, but if you get lots of messages about > issues watching from specific versions try that. > > > If you have only one node, what is the benefit of a high availability master? > The cluster will keep happily running if a single master is temporarily > unavailable, but whenever the single node goes away the applications will be > effectively down. > > > /MR > > Just for testing, I m using 1 node cluster can you elaborate setting up the load balancer, I very new to kuberentes
-- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.