RethinkDB Cluster Helm Chart

Prerequisites Details

• Kubernetes 1.5+ with Beta APIs enabled.
• PV support on the underlying infrastructure.

StatefulSet Details

• https://kubernetes.io/docs/concepts/abstractions/controllers/statefulsets/

StatefulSet Caveats

• https://kubernetes.io/docs/concepts/abstractions/controllers/statefulsets/#limitations

Acknowledgment of Previous Works

This is a simple fork of the deprecated official helm chart for RethinkDB which was originally contributed by https://github.com/meenie.

That work too heavily borrowed and extended code (peer discovery and probe) from the following project to build the Helm Chart and Docker image: https://github.com/rosskukulinski/kubernetes-rethinkdb-cluster

Chart Details

This chart implements a dynamically scalable RethinkDB Cluster using Kubernetes StatefulSets.

Installing the Chart

To install the chart with the release name my-release:

$ helm install --name my-release pozetron/rethinkdb

Configuration

The following table lists the configurable parameters of the rethinkdb chart and their default values.

Parameter	Description	Default
image.name	Custom RethinkDB image name for auto-joining and probe	pozetroninc/rethinkdb-cluster
image.tag	Custom RethinkDB image tag	0.1.0
image.pullPolicy	Custom RethinkDB image pull policy	IfNotPresent
cluster.replicas	Number of RethinkDB Cluster replicas	3
cluster.resources	Resource configuration for each RethinkDB Cluster Pod	{}
cluster.podAnnotations	Annotations to be added to RethinkDB Cluster Pods	{}
cluster.service.annotations	Annotations to be added to RethinkDB Cluster Service	{}
cluster.storageClass.enabled	If true, create a StorageClass for the cluster. Note: You must set a provisioner	false
cluster.storageClass.provisioner	Provisioner definition for StorageClass	undefined
cluster.storageClass.parameters	Parameters for StorageClass	undefined
cluster.persistentVolume.enabled	If true, persistent volume claims are created	true
cluster.persistentVolume.storageClass	Persistent volume storage class	default
cluster.persistentVolume.accessMode	Persistent volume access modes	[ReadWriteOnce]
cluster.persistentVolume.size	Persistent volume size	1Gi
cluster.persistentVolume.annotations	Persistent volume annotations	{}
cluster.rethinkCacheSize	RethinkDB cache-size value in MB	100
proxy.replicas	Number of RethinkDB Proxy replicas	1
proxy.resources	Resource configuration for each RethinkDB Proxy Pod	{}
proxy.podAnnotations	Annotations to be added to RethinkDB Proxy Pods	{}
proxy.service.type	RethinkDB Proxy Service Type	ClusterIP
proxy.service.annotations	Annotations to be added to RethinkDB Cluster Service	{}
proxy.service.clusterIP	Internal controller proxy service IP	""
proxy.service.externalIPs	Controller service external IP addresses	[]
proxy.service.loadBalancerIP	IP address to assign to load balancer (if supported)	""
proxy.service.loadBalancerSourceRanges	List of IP CIDRs allowed access to load balancer (if supported)	[]
proxy.driverTLS.enabled	Should RethinkDB Proxy TLS be enabled. Note: If enabled, you must set a key and cert	false
proxy.driverTLS.key	RSA Private Key	undefined
proxy.driverTLS.cert	Certificate	undefined
ports.cluster	RethinkDB Cluster Port	29015
ports.driver	RethinkDB Driver Port	28015
ports.admin	RethinkDB Admin Port	8080
rethinkdbPassword	Password for the RethinkDB Admin user	rethinkdb

Specify each parameter using the --set key=value[,key=value] argument to helm install.

Alternatively, a YAML file that specifies the values for the parameters can be provided while installing the chart. For example,

$ helm install --name my-release -f values.yaml pozetron/rethinkdb

Tip: You can use the default values.yaml

Important: Admin Password Management

The initial admin password is set by the config value rethinkdbPassword. This value is also used by the probe which periodically checks if the RethinkDB Cluster and Proxy are still running. If you change the RethinkDB admin password via a query (i.e. r.db('rethinkdb').table('users').update({password: 'new-password'})) this will cause the probe to fail which then restarts the pods over and over. To stablize the cluster, you also need to use helm upgrade to update the password in the Kubernetes Secrets storage by doing:

$ helm upgrade --set rethinkdbPassword=new-password my-release pozetron/rethinkdb

Opening Up the RethinkDB Admin Console

The admin port is not available outside of the cluster for security reasons. The only way to access the admin console is to use a Kubernetes Proxy. To open up the admin console:

$ kubectl proxy
Starting to serve on 127.0.0.1:8001

Then use the following URL: http://localhost:8001/api/v1/namespaces/NAMESPACE/services/RELEASE_NAME-rethinkdb-admin/proxy Make sure a replace NAMEPSPACE with the correct namespace and RELEASE_NAME that was used when installing the chart.

And then open up your browser to http://localhost:8080 and you should see the admin console

Cleanup orphaned Persistent Volumes

Deleting a StatefulSet will not delete associated Persistent Volumes.

Do the following after deleting the chart release to clean up orphaned Persistent Volumes.

$ kubectl delete pvc -l release=my-release

Failover

If any RethinkDB server fails it gets re-joined eventually. You can test the scenario by killing process of one of the pods:

$ kubectl get pods -l release=my-release
NAME                                          READY     STATUS    RESTARTS   AGE
my-release-rethinkdb-cluster-0                1/1       Running   0          1m
my-release-rethinkdb-cluster-1                1/1       Running   0          2m
my-release-rethinkdb-cluster-2                1/1       Running   0          2m
my-release-rethinkdb-proxy-2517940628-81dxd   1/1       Running   1          1m

$ kubectl exec -it my-release-rethinkdb-cluster-0 -- ps aux | grep 'rethinkdb'
root         7  0.1  2.1 233496 43408 ?        Ssl  16:56   0:00 rethinkdb --ser
root        26  0.0  0.4 146948  8204 ?        S    16:56   0:00 rethinkdb --ser
root       100  0.0  0.7 157192 16060 ?        S    16:56   0:00 rethinkdb --ser

$ kubectl exec -it my-release-rethinkdb-cluster-0 -- kill 7

Scaling

Scaling should be managed by helm upgrade, which is the recommended way. Example:

$ helm upgrade --set cluster.replicas=4 my-release pozetron/rethinkdb