From d228222fa6c9c5e3e2367ca69d87ae508d48aeb8 Mon Sep 17 00:00:00 2001
From: Jerome Petazzoni <jerome.petazzoni@gmail.com>
Date: Sun, 8 Apr 2018 17:59:42 -0500
Subject: [PATCH] Reword headless services

Hopefully this explains better the use of headless services.
I also added a slide about endpoints, with a couple of simple
commands to show them.
---
 slides/kube/kubectlexpose.md | 85 +++++++++++++++++++++++++++---------
 1 file changed, 65 insertions(+), 20 deletions(-)

diff --git a/slides/kube/kubectlexpose.md b/slides/kube/kubectlexpose.md
index fb9b30ce..654ff498 100644
--- a/slides/kube/kubectlexpose.md
+++ b/slides/kube/kubectlexpose.md
@@ -53,26 +53,6 @@ The `LoadBalancer` type is currently only available on AWS, Azure, and GCE.
 
 ---
 
-## Headless services
-
-- A "headless service" is a service that doesn't have a virtual IP address
-
-- Since there is no virtual IP address, there is no load balancer either
-
-- This is useful if we want to connect directly to the pods
-
-- We can discover the pod IP addresses:
-
-  - by querying the Kubernetes API
-  - by querying the internal DNS (if we use `kube-dns`)
-  - with other custom mechanisms (e.g. running Consul, Zookeeper...)
-
-- A headless service is obtained by setting the `clusterIP` field to `None`
-
-  (Either with `--cluster-ip=None`, or by providing a custom YAML)
-
----
-
 ## Running containers with open ports
 
 - Since `ping` doesn't have anything to connect to, we'll have to run something else
@@ -158,3 +138,68 @@ Note: please DO NOT call the service `search`. It would collide with the TLD.
 --
 
 Our requests are load balanced across multiple pods.
+
+---
+
+class: extra-details
+
+## If we don't need a load balancer
+
+- Sometimes, we want to access our scaled services directly:
+
+  - if we want to save a tiny little bit of latency (typically less than 1ms)
+
+  - if we need to connect over arbitrary ports (instead of a few fixed ones)
+
+  - if we need to communicate over another protocol than UDP or TCP
+
+  - if we want to decide how to balance the requests client-side
+
+  - ...
+
+- In that case, we can use a "headless service"
+
+---
+
+class: extra-details
+
+## Headless services
+
+- A headless service is obtained by setting the `clusterIP` field to `None`
+
+  (Either with `--cluster-ip=None`, or by providing a custom YAML)
+
+- As a result, the service doesn't have a virtual IP address
+
+- Since there is no virtual IP address, there is no load balancer either
+
+- `kube-dns` will return the pods' IP addresses as multiple `A` records
+
+- This gives us an easy way to discover all the replicas for a deployment
+
+---
+
+class: extra-details
+
+## Services and endpoints
+
+- A service has a number of "endpoints"
+
+- Each endpoint is a host + port where the service is available
+
+- The endpoints are maintained and updated automatically by Kubernetes
+
+.exercise[
+
+- Check the endpoints that Kubernetes has associated with our `elastic` service:
+  ```bash
+  kubectl get service elastic -o wide
+  kubectl get endpoints elastic
+  ```
+
+- Compare with the pods for the `elastic` service:
+  ```bash
+  kubectl get pods -l run=elastic -o wide
+  ```
+
+]