Opni Opensearch Operator

The Opni Opensearch Operator is a Kubernetes operator designed to deploy and manage an Opensearch cluster

Features

The Opensearch operator currently provides the following features:

• Deploy an Opensearch cluster with separate master, data, and client nodes.
• Manage certificates for the cluster, and rotate certificates when they are about to expire.
• Perform a rolling upgrade on the Opensearch cluster when a new version is applied in the custom resource.

Getting started

To install the operator into your Kubernetes cluster you can use the kustomize manifests:

kubectl create -k config/default

Custom Resources

OpensearchCluster

This controls the deployment of an Opensearch cluster.

example.yaml

apiVersion: opensearch.opni.io/v1beta1
kind: OpensearchCluster
metadata:
  name: example
  namespace: opni
spec:
  version: 1.1.0
  master:
    replicas: 3
    resources:
      requests:
        memory: 1Gi
      limits:
        memory: 1Gi
  data:
    replicas: 2
    resources:
      requests:
        memory: 2Gi
      limits:
        memory: 2Gi

OpensearchSpec

Field	Required	Type	Description
version	No	string	Version of the Opendistro image to use. Defaults to latest
defaultRepo	No	string	Image repo to use for Opensearch images. Defaults to docker.io/amazon
image	No	ImageSpec	Explicit configuration for the Opensearch image
configSecret	No	LocalObjectReference	Reference to a secret containing logging.yml with the Opensearch logging config, and opensearch.yml with the Opensearch config
authConfigSecret	No	LocalObjectReference	Reference to a secret containing config.yml with the Opensearch auth config
adminPasswordFrom	No	SecretKeySelector	Secret key selector pointing to the admin password to use. If not set a password will be autogenerated and stored in the password key in a secret called opni-es-password
master	ElasticWorkloadOptions	Configuration for the Elasticsearch master node StatefulSet
data	ElasticWorkloadOptions	Configuration for the Elasticsearch data node StatefulSet
client	ElasticWorkloadOptions	Configuration for the Elasticsearch client node Deployment

ElasticWorkloadOptions

Field	Required	Type	Description
replicas	No	int	Number of replicas to deploy. Defaults to 1
resources	No	ResourceRequirements	Resources for the workload. The resources are used to calculate the Java memory options. If a memory limit exists this is used, if there is a memory request but no limit that is used, otherwise the Java optioms default to a heap size of 512MB
affinity	No	Affinity	Affinity settings for the workload pods
nodeSelector	No	map	NodeSelector for the workload pods. If this exists it will override the globalNodeSelector
tolerations	No	Toleration array	Tolerations for the workload. These will be combined with the globalTolerations (if any)
persistence	No	PersistenceSpec	Configure the persistence for Elasticsearch data
extraEnvVars	No	array EnvVar	Additional Env Vars to add to the pods

ImageSpec

Field	Required	Type	Description
image	No	string	Explicit CRI image to use
imagePullPolicy	No	string	Image pull policy. One of Always, Never, IfNotPresent. Defaults to IfNotPresent
imagePullSecrets	No	LocalObjectReference array	List of secrets in the same namespace to use for pulling the image

PersistenceSpec

Field	Required	Type	Description
enabled	No	bool	Whether persistent storage is enabled. Defaults to false
storageClassName	No	string	If persistent storage is enabled, the name of the StorageClass to use. If not set will use the default StorageClass
accessModes	No	string array	An array of the access modes the volume supports
request	No	string	The size of the volume to request. Defaults to 10Gi

Dashboards

This controls the deployment of Opensearch Dashboards.

example.yaml

apiVersion: opensearch.opni.io/v1beta1
kind: Dashboards
metadata:
  name: example
  namespace: opni
spec:
  version: 1.1.0
  opensearch:
    name: example

DashboardsSpec

Field	Required	Type	Description
version	No	string	Version of the Opendistro Dashboards image to use. Defaults to latest
defaultRepo	No	string	Image repo to use for Opensearch Dashboards images. Defaults to docker.io/amazon
image	No	ImageSpec	Explicit configuration for the Opensearch Dashboards image
replicas	No	int	Number of replicas to deploy. Defaults to 1
resources	No	ResourceRequirements	Resources for the workload. The resources are used to calculate the Java memory options. If a memory limit exists this is used, if there is a memory request but no limit that is used, otherwise the Java optioms default to a heap size of 512MB
affinity	No	Affinity	Affinity settings for the workload pods
nodeSelector	No	map	NodeSelector for the workload pods. If this exists it will override the globalNodeSelector
tolerations	No	Toleration array	Tolerations for the workload. These will be combined with the globalTolerations (if any)
opensearch	No	LocalObjectReference	Reference to an existing OpensearchCluster to point the Dashboards deployment at. Must be in the same namespace
opensearchConfig	No	OpensearchConfigSpec	Configuration for an external Opensearch cluster
tlsSecret	No	LocalObjectReference	A TLS secret containing the cert and key to use for Dashboards SSL. If the opensearch field is preset this isn't required as it will reuse the Opensearch HTTP CA to generate a cert

OpensearchConfigSpec

Field	Required	Type	Description
url	Yes	string	Endpoint for the Opensearch cluster
username	Yes	string	Username to connect to the cluster with
passwordFrom	Yes	SecretKeySelector	Secret key which contains the password for the cluster
verifySSL	No	bool	Whether to use strict SSL checking. Defaults to true