Skip to content

Install Gluu Flex on GKE#

System Requirements#

The resources may be set minimally to the below:

  • 8-13 GB RAM based on the services deployed
  • 8-11 CPU cores based on the services deployed
  • 50GB hard-disk

Use the listing below for a detailed estimation of the minimum required resources. The table contains the default resources recommendation per service. Depending on the use of each service the resources need may be increase or decrease.

Service CPU Unit RAM Disk Space Processor Type Required
Auth server 2.5 2.5GB N/A 64 Bit Yes
fido2 0.5 0.5GB N/A 64 Bit No
scim 1 1GB N/A 64 Bit No
config - job 0.3 0.3GB N/A 64 Bit Yes on fresh installs
persistence - job 0.3 0.3GB N/A 64 Bit Yes on fresh installs
nginx 1 1GB N/A 64 Bit Yes if ALB/Istio not used
auth-key-rotation 0.3 0.3GB N/A 64 Bit No [Strongly recommended]
config-api 1 1GB N/A 64 Bit No
casa 0.5 0.5GB N/A 64 Bit No
admin-ui 2 2GB N/A 64 Bit No
link 0.5 1GB N/A 64 Bit No
saml 0.5 1GB N/A 64 Bit No

Releases of images are in style 1.0.0-beta.0, 1.0.0-0

Initial Setup#

  1. Before initiating the setup, please obtain an SSA for Flex trial, after which you will issued a JWT.

  2. Enable GKE API if not enabled yet.

  3. If you are using Cloud Shell, you can skip to step 7.

  4. Install gcloud.

  5. Install kubectl using gcloud components install kubectl command.

  6. Install Helm3.

  7. Create cluster using a command such as the following example:

    gcloud container clusters create gluu-cluster --num-nodes 2 --machine-type e2-standard-4 --zone us-west1-a
    You can adjust num-nodes and machine-type as per your desired cluster size

  8. Create gluu namespace where our resources will reside 

    kubectl create namespace gluu

Gluu Flex Installation using Helm#

  1. Install Nginx-Ingress, if you are not using Istio ingress

    helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
helm repo add stable https://charts.helm.sh/stable
helm repo update
helm install nginx ingress-nginx/ingress-nginx

  2. Create a file named override.yaml and add changes as per your desired configuration:

    • FQDN/domain is not registered:

      Get the Loadbalancer IP: 

      kubectl get svc nginx-ingress-nginx-controller --output jsonpath='{.status.loadBalancer.ingress[0].ip}'

      Add the following yaml snippet to your override.yaml file:

      global:
    lbIp: #Add the Loadbalance IP from the previous command
    isFqdnRegistered: false

    • FQDN/domain is registered:

      Add the following yaml snippet to your override.yaml file:

      global:
    lbIp: #Add the LoadBalancer IP from the previous command
    isFqdnRegistered: true
    fqdn: demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu
nginx-ingress:
  ingress:
      path: /
      hosts:
      - demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu
      tls:
      - secretName: tls-certificate
        hosts:
        - demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu

    • Couchbase for pesistence storage

      Add the following yaml snippet to your override.yaml file:

      global:
  cnPersistenceType: couchbase

config:
  configmap:
    # The prefix of couchbase buckets. This helps with separation in between different environments and allows for the same couchbase cluster to be used by different setups of Flex.
    cnCouchbaseBucketPrefix: jans
    # -- Couchbase certificate authority string. This must be encoded using base64. This can also be found in your couchbase UI Security > Root Certificate. In mTLS setups this is not required.
    cnCouchbaseCrt: SWFtTm90YVNlcnZpY2VBY2NvdW50Q2hhbmdlTWV0b09uZQo=
    # -- The number of replicas per index created. Please note that the number of index nodes must be one greater than the number of index replicas. That means if your couchbase cluster only has 2 index nodes you cannot place the number of replicas to be higher than 1.
    cnCouchbaseIndexNumReplica: 0
    # -- Couchbase password for the restricted user config.configmap.cnCouchbaseUser that is often used inside the services. The password must contain one digit, one uppercase letter, one lower case letter and one symbol
    cnCouchbasePassword: P@ssw0rd
    # -- The Couchbase super user (admin) username. This user is used during initialization only.
    cnCouchbaseSuperUser: admin
    # -- Couchbase password for the superuser config.configmap.cnCouchbaseSuperUser that is used during the initialization process. The password must contain one digit, one uppercase letter, one lower case letter and one symbol
    cnCouchbaseSuperUserPassword: Test1234#
    # -- Couchbase URL. This should be in FQDN format for either remote or local Couchbase clusters. The address can be an internal address inside the kubernetes cluster
    cnCouchbaseUrl: cbjanssen.default.svc.cluster.local
    # -- Couchbase restricted user
    cnCouchbaseUser: janssen

    • PostgreSQL for persistence storage

      In a production environment, a production grade PostgreSQL server should be used such as Cloud SQL

      For testing purposes, you can deploy it on the GKE cluster using the following command:

      helm install my-release --set auth.postgresPassword=Test1234#,auth.database=gluu -n gluu oci://registry-1.docker.io/bitnamicharts/postgresql

      Add the following yaml snippet to your override.yaml file:

      global:
  cnPersistenceType: sql
config:
  configmap:
    cnSqlDbName: gluu
    cnSqlDbPort: 5432
    cnSqlDbDialect: pgsql
    cnSqlDbHost: my-release-postgresql.gluu.svc
    cnSqlDbUser: postgres
    cnSqlDbTimezone: UTC
    cnSqldbUserPassword: Test1234#

    • MySQL for persistence storage

      In a production environment, a production grade MySQL server should be used such as Cloud SQL

      For testing purposes, you can deploy it on the GKE cluster using the following command:

      helm install my-release --set auth.rootPassword=Test1234#,auth.database=gluu -n gluu oci://registry-1.docker.io/bitnamicharts/mysql

      Add the following yaml snippet to your override.yaml file:

      global:
  cnPersistenceType: sql
config:
  configmap:
    cnSqlDbName: gluu
    cnSqlDbPort: 3306
    cnSqlDbDialect: mysql
    cnSqlDbHost: my-release-mysql.gluu.svc
    cnSqlDbUser: root
    cnSqlDbTimezone: UTC
    cnSqldbUserPassword: Test1234#

      So if your desired configuration has FQDN and MySQL, the final override.yaml file will look something like that:

      global:
  cnPersistenceType: sql
  lbIp: "" #Add the LoadBalancer IP from previous command
  isFqdnRegistered: true
  fqdn: demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu
nginx-ingress:
  ingress:
      path: /
      hosts:
      - demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu
      tls:
      - secretName: tls-certificate
        hosts:
        - demoexample.gluu.org #CHANGE-THIS to the FQDN used for Gluu  
config:
  configmap:
    cnSqlDbName: gluu
    cnSqlDbPort: 3306
    cnSqlDbDialect: mysql
    cnSqlDbHost: my-release-mysql.gluu.svc
    cnSqlDbUser: root
    cnSqlDbTimezone: UTC
    cnSqldbUserPassword: Test1234#

  3. Install Gluu Flex

    After finishing all the tweaks to the override.yaml file, we can use it to install gluu flex.

    helm repo add gluu-flex https://docs.gluu.org/charts
helm repo update
helm install gluu gluu-flex/gluu -n gluu -f override.yaml

Configure Gluu Flex#

You can use the Janssen TUI to configure Flex components. The TUI calls the Config API to perform ad hoc configuration.

Last update: 2024-10-24
Created: 2022-09-22