Janusgraph is an open source graph database that is highly scalable and optimized for storing and querying large graph. In this article, we will be looking into how to Helm charts for Janusgraph which:

Use Cassandra as storage
Use ElasticSearch for indexing

Note: We will be running in Window docker desktop environment.

Source code: GitHub

Whats is Helm

Helm is a package manager for k8s that equivalent to apt or yum on linux. It deploys chart(~package application).

Helm chart folder structure

Below is the folder structure for our Janusgraph Helm’s chart:

-> janusgraph-helm
   -> templates
      -> janusgraph.yaml
   -> Chart.yaml
   -> requirements.yaml
   -> values.yaml

Chart.yaml

A YAML file which contains the information of the chart. We will describe the chart as below:

apiVersion: v1
name: my-janusgraph
version: 0.1.0
appVersion: 1.0
description: A simple customized janusgraph chart that use cassandara and elastic search

requirements.yaml

A YAML file which lists out the dependencies for the chart. We will be using Cassandra and Elasticsearch form Bitnami.

dependencies:
- name: cassandra
  version: 5.6.1
  repository: "https://charts.bitnami.com/bitnami"
  alias: cassandra
- name: elasticsearch
  version: 12.6.1
  repository: "https://charts.bitnami.com/bitnami"
  alias: elasticsearch

templates

Template directory is the place we place our k8s manifest file to describe how we want the cluster to be deployed. By default, Helm use Go templating engine.

We will be creating a Deployment and Service for Janusgraph.

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: janusgraph
  name: janusgraph
spec:
  replicas: 1
  selector:
    matchLabels:
      app: janusgraph
  template:
    metadata:
      labels:
        app: janusgraph
    spec:
      initContainers:
      - name: check-elasticsearch
        image: busybox
        command: ['sh', '-c', 'until [ "$(wget -q -O - janusgraph-helm-elasticsearch-master:9200/_cat/nodes | wc -l)" = "" ];
                  do echo "Waiting for Elasticsearch cluster..."; sleep 10; done;
                  echo "Elasticsearch cluster is ready, launching JanusGraph Server..."']
      containers:
      - image: janusgraph/janusgraph:0.5.2
        name: janusgraph
        env:
          - name: JANUS_PROPS_TEMPLATE
            value: cql-es
          - name: janusgraph.storage.hostname
            value: janusgraph-helm-cassandra
          - name: janusgraph.storage.username
            value: cassandra
          - name: janusgraph.storage.password
            value: password
          - name: janusgraph.index.search.hostname
            value: janusgraph-helm-elasticsearch-master

Notes:

Janusgraph version 0.5.2 will be used.
It will only start initializing when elasticsearch is ready.
We have also specified cassandra and elasticsearch hostname in the containers.env.

apiVersion: v1
kind: Service
metadata:
  labels:
    app: janusgraph
  name: janusgraph
spec:
  ports:
  - port: 8182
    name: janusgraph
    protocol: TCP
    targetPort: 8182
  selector:
    app: janusgraph
  type: ClusterIP

values.yaml

The configuration value for the chart.

global:
  storageClass: hostpath
cassandra:
  dbUser:
    user: cassandra
    password: password
elasticsearch:
  master:
    replicas: 1
  coordinating:
    replicas: 1
  data:
    replicas: 1

Provision Janusgraph Cluster

Deploy Janusgraph Cluster

Before deploying Janusgraph we will need to make sure Helm pulls the required dependencies.

helm dependency update

After the dependency fulfilled, we can run the following command to deploy Janusgraph cluster to k8s environment.

helm install janusgraph-helm .

Validate Environment

Run kubectl get pods to check whether environment is up. Example:

 ❯ kubectl get pods
 NAME                                                              READY   STATUS    RESTARTS   AGE
 janusgraph-9bf995c79-lhmtg                                        1/1     Running   2          6h31m
 janusgraph-helm-cassandra-0                                       1/1     Running   2          6h31m
 janusgraph-helm-elasticsearch-coordinating-only-549d8b75c-g6mqz   1/1     Running   2          6h31m
 janusgraph-helm-elasticsearch-data-0                              1/1     Running   2          6h31m
 janusgraph-helm-elasticsearch-master-0                            1/1     Running   2          6h31m

Next, we will bring up the gremlin interactive console.
We know that our janusgraph pod named janusgraph-9bf995c79-lhmtg , so run the following command to bring up the janusgraph command.
```
1
 kubectl exec -it janusgraph-9bf995c79-lhmtg -- bash
```

Execute bin/gremlin.sh and you should

 root@janusgraph-9bf995c79-lhmtg:/opt/janusgraph# bin/gremlin.sh
 Nov 15, 2020 9:00:41 AM java.util.prefs.FileSystemPreferences$1 run
 INFO: Created user preferences directory.
    
          \,,,/
          (o o)
 -----oOOo-(3)-oOOo-----
 SLF4J: Class path contains multiple SLF4J bindings.
 SLF4J: Found binding in [jar:file:/opt/janusgraph/lib/slf4j-log4j12-1.7.12.jar!/org/slf4j/impl/StaticLoggerBinder.class]
 SLF4J: Found binding in [jar:file:/opt/janusgraph/lib/logback-classic-1.1.3.jar!/org/slf4j/impl/StaticLoggerBinder.class]
 SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
 SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
 plugin activated: tinkerpop.server
 plugin activated: tinkerpop.tinkergraph
 09:00:43 WARN  org.apache.hadoop.util.NativeCodeLoader  - Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
 plugin activated: tinkerpop.hadoop
 plugin activated: tinkerpop.spark
 plugin activated: tinkerpop.utilities
 plugin activated: janusgraph.imports
 gremlin>

Connect to the janusgraph gremlin server.

 gremlin> :remote connect tinkerpop.server conf/remote.yaml
 ==>Configured localhost/127.0.0.1:8182

Point the gremlin console from local to janusgraph gremlin server.

 gremlin> :remote console
 ==>All scripts will now be sent to Gremlin Server - [localhost/127.0.0.1:8182] - type ':remote console' to return to local mode

Run node count to validate.

 gremlin> g.V().count()
 ==>0

Deploy Janusgraph on Kubernetes with Helm