M
M
MherArsh2018-08-01 16:13:34
Kubernetes
MherArsh, 2018-08-01 16:13:34

Failsafe Ingress, how does it work? how to implement?

The situation is as follows:
Assuming we have several nodes (node1, node2) on which we have an IngressController, in theory, ingress is a single access point ... if we have 2 nodes and 2 replicas of ingress, my my.example.com on which the address should indicate that when one node falls, access to the service was?
I now have a cluster running on Ranche 1.x, where ingress rises on the first node (node1) and in my.example.com I write the address of node1, after adding the second node (node2) even if I make replication clear that this will not help if node1 will fall because my.example.com has its ip registered.
Question: how can I make it so that there is one address / entry point (I don’t know how to formulate it correctly), so that the drinking of one node does not affect the availability of the service?
Thanks in advance.

Answer the question

In order to leave comments, you need to log in

1 answer(s)
M
Max, 2018-08-28
@MaxDukov

look here . Apparently, it’s quite beautiful only through virtual IP (I will be VERY happy if they poke my nose into a more elegant solution).
if a half-hearted solution - then either through DNS round robin, which is never HA, or through an additional external balancer - which also leads us to a single point of failure.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question