kube-proxy "watches" endpoints-- this means apiserver pushes changes (add/update/delete) in real time.
Endpoints are maintained by an imaginatively named "endpoints controller"-- it watches pods and services from apiserver and writes the endpoints from ready pods back to apiserver. Pod statuses (liveness & readiness) are detected by kubelet and written to apiserver as they change. Replication controller is not involved in this at all, it just makes sure the right number of pods are running. On Sat, May 21, 2016 at 12:31 PM, kant kodali <[email protected]> wrote: > How does Kube-Proxy know one of the service pod is down and that it needs > to connect to the new one? I looked into some of the kube-proxy code and it > looks like it constantly listens for change in endpoints but what tells the > kube-proxy that there is a new end point? API server? If so how does API > server know that there is a new endpoint? Is it the replication controller > which make sure's that there are specified number of pods running at any > given time(similar to linux monit) and then the replication controller > sends message to API server? > > -- > You received this message because you are subscribed to the Google Groups > "Containers at Google" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/google-containers. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Containers at Google" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/google-containers. For more options, visit https://groups.google.com/d/optout.
