Now that the Mcrouter Operator is deployed, we can apply the kind: McRouter Custom Resource.

The Mcrouter Custom Resource contains a number of configurable options in the spec section.

# The image to run for the `mcrouter` Deployment.
mcrouter_image: devan2502/mcrouter:latest

# The port Mcrouter will run on.
mcrouter_port: 5000

# The image to run for the `memcached` StatefulSet.
memcached_image: memcached:1.5-alpine

# The size of the memcached pool.
memcached_pool_size: 3

# The port Memcached will run on.
memcached_port: 11211

# The memcached pool can be 'sharded' or 'replicated'.
# 'sharded' uses a key hashing algorithm to distribute Memcached sets and gets among Memcached Pods; this means a key foo may always go to pod A, while the key bar always goes to pod B.
# 'replicated' sends all Memcached sets to all Memcached pods, and distributes gets randomly.
pool_setup: replicated

# Set to '/var/mcrouter/fifos' to debug mcrouter with mcpiper.
debug_fifo_root: ''

Let's create an instance of Mcrouter that will create a replicated Memcached pool size of 2. This will create a Deployment consisting of 1 pod (image: devan2502/mcrouter:latest ) as well as a Memcached StatefulSet consists of 2 memcached pods (memcached_image: memcached:1.5-alpine).

apiVersion: mcrouter.example.com/v1alpha3
kind: Mcrouter
metadata:
  name: mcrouter
spec:
  mcrouter_image: devan2502/mcrouter:latest
  mcrouter_port: 5000
  memcached_image: memcached:1.5-alpine
  memcached_pool_size: 2
  memcached_port: 11211
  pool_setup: replicated # or 'sharded'
  debug_fifo_root: '' # set to '/var/mcrouter/fifos' for mcpiper

Before applying the Custom Resource (CR), open up a new terminal window and follow the logs for the ansible container. This will expose the standard Ansible stdout logs and we should see activity once the CR is created.

oc logs deploy/mcrouter-operator -c ansible -f

Run the following to apply the CR:

oc apply -f https://raw.githubusercontent.com/geerlingguy/mcrouter-operator/master/deploy/crds/mcrouter_v1alpha3_mcrouter_cr.yaml

Verify the Mcrouter CR now exists:

oc get mcrouter

Verify the mcrouter deployment pod:

oc get pod -l name=mcrouter-operator

The mcrouter Deployment Pod is exposed via a Service listening on the default memcached port, 11211. Mcrouter will connect to the pool via the backend memcached instances residing in the pool via this Service.

oc get svc mcrouter-memcached

Verify the two mcrouter memcache instances are running:

oc get pods -l app=mcrouter-cache

We can now verify connectivity to our Memcached pool via Mcrouter.

Let's begin by exposing the Mcrouter Pod as a Service accessible to the Cluster.

oc expose deploy/mcrouter --name mcrouter --type=ClusterIP --target-port=5000

Create a temporary telnet Pod and connect to Mrouter:

oc run -it --rm telnet --image=jess/telnet --restart=Never mcrouter 5000

After a few seconds you will see a message If you don't see a command prompt, try pressing enter. Do not press enter. Run the following commands to insert a key/value and exit:

set ansible 0 0 8
operator
get ansible
quit

We can also utilize the mcperf tool for measuring Memcached server performance.

This command will create 1000 connections to Mcrouter. The connections are created at the rate of 1000 conns/sec and on every connection it sends 10 set requests at the rate of 1000 reqs/sec with the item sizes derived from a uniform distribution in the interval of 1 to 16 bytes:

oc run -it --rm mcperf --image=quay.io/redhat/mcperf --restart=Never -- -s mcrouter -p 5000 --linger=0 --timeout=5 --conn-rate=1000 --call-rate=1000 --num-calls=10 --num-conns=1000 --sizes=u1,16

We should be able to see a significant spike in CPU/Memory usage from running the above command. You can check it out by selecting the Dashboard tab which will open the OpenShift dashboard. Log in with username: kubeadmin, password: admin, and then select Monitoring --> Dashboards. Also be sure to run the following command:

oc adm policy add-cluster-role-to-user cluster-admin kubeadmin

One simple way to verify Mcrouter Operator functionality is by changing the memcached_pool_size in the Mcrouter Custom Resource. Run the folllowing command to bump the pool size to 3 and then observe what happens in the cluster:

oc patch mcrouter mcrouter  --type='json' -p '[{"op": "replace", "path": "/spec/memcached_pool_size", "value":3}]'

After a few seconds, a new Memcached Pod should be added to the pool:

oc get pods -l app=mcrouter-cache

You should now see the additional Memcached instances reflected in Mcrouter's configuration string.

oc describe pod -l app=mcrouter | grep Command -A2

Thanks to the embedded StatefulSet controller our Memcached pool will handle creation of a new Pod in the event of failure.

Fetch one of Mcrouter's Memcached Pods from the pool:

MCROUTER_CACHE_0=`oc get pods -l app=mcrouter-cache -o jsonpath={$.items[0].metadata.name}`
echo $MCROUTER_CACHE_0

Delete the Pod:

oc delete pod $MCROUTER_CACHE_0

The Pod should respawn:

oc get pods -l app=mcrouter-cache