Enterprise Java

Apache Ignite and Spring on your Kubernetes Cluster Part 2: Kubernetes deployment

Previously we have been successful on creating our first Spring boot Application powered by Apache Ignite.

On this blog we shall focus on what is needed to be done on the Kubernetes side in order to be able to spin up our application.

As described on a previous blog we need to have our Kubernetes RBAC policies in place.

We need a role, a service account and the binding.

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: job-cache
rules:
  - apiGroups:
    - ""
    resources:
    - pods
    - endpoints
    verbs:
    - get
    - list
    - watch
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: job-cache
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  creationTimestamp: 2020-03-07T22:23:50Z
  name: job-cache
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: job-cache
subjects:
  - kind: ServiceAccount
    name: job-cache
    namespace: "default"

Our service account will be the job cache. This means that we should use the job-cache service account for our Ignite based workloads.

The next step is to create the deployment. The configuration would not be very different from statefulset as explained in a previous post.

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
apiVersion: apps/v1
kind: Deployment
metadata:
  name: job-api-deployment
  labels:
    app: job-api
spec:
  replicas: 2
  selector:
    matchLabels:
      app: job-api
  template:
    metadata:
      labels:
        app: job-api
    spec:
      containers:
        - name: job-api
          image: job-api:1.0
          env:
            - name: IGNITE_QUIET
              value: "false"
            - name: IGNITE_CACHE_CLIENT
              value: "false"
          ports:
            - containerPort: 11211
              protocol: TCP
            - containerPort: 47100
              protocol: TCP
            - containerPort: 47500
              protocol: TCP
            - containerPort: 49112
              protocol: TCP
            - containerPort: 10800
              protocol: TCP
            - containerPort: 8080
              protocol: TCP
            - containerPort: 10900
              protocol: TCP
      serviceAccount: job-cache
      serviceAccountName: job-cache

This is simpler since the Ignite configuration has been done through Java code.
The image that you see is supposed to be your dockerised Java application we worked before.
The next big step is to define the service. I will not use one service for all. Instead I would create a service for the cache and a service for our api in order to be used as an api.

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
apiVersion: v1
kind: Service
metadata:
  labels:
    app: job-cache
  name: job-cache
spec:
  ports:
    - name: jdbc
      port: 11211
      protocol: TCP
      targetPort: 11211
    - name: spi-communication
      port: 47100
      protocol: TCP
      targetPort: 47100
    - name: spi-discovery
      port: 47500
      protocol: TCP
      targetPort: 47500
    - name: jmx
      port: 49112
      protocol: TCP
      targetPort: 49112
    - name: sql
      port: 10800
      protocol: TCP
      targetPort: 10800
    - name: rest
      port: 8080
      protocol: TCP
      targetPort: 8080
    - name: thin-clients
      port: 10900
      protocol: TCP
      targetPort: 10900
  selector:
    app: job-api
  type: ClusterIP

Without getting into kubernetes details, the Ignite nodes shall synchronize using the job-cache internal dns. So we shall use kubernetes internal dns capabilities to communicate with the Ignite cluster.

The next step is to create the service for the actual job api application.

01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
apiVersion: v1
kind: Service
metadata:
  labels:
    app: job-api
  name: job-api
spec:
  ports:
    - name: rest-api
      port: 80
      protocol: TCP
      targetPort: 8080
  selector:
    app: job-api
  sessionAffinity: None
  type: ClusterIP

Οn the following blog we shall apply our configurations to kubernetes and test our codebase.

Published on Java Code Geeks with permission by Emmanouil Gkatziouras, partner at our JCG program. See the original article here: Apache Ignite and Spring on your Kubernetes Cluster Part 2: Kubernetes deployment

Opinions expressed by Java Code Geeks contributors are their own.

Emmanouil Gkatziouras

He is a versatile software engineer with experience in a wide variety of applications/services.He is enthusiastic about new projects, embracing new technologies, and getting to know people in the field of software.
Subscribe
Notify of
guest

This site uses Akismet to reduce spam. Learn how your comment data is processed.

0 Comments
Inline Feedbacks
View all comments
Back to top button