keycloak-scim/openshift/topics/tutorials.adoc

[[{project_name}-Binary-Builds-Tutorial]]
=== Example Workflow: Creating OpenShift Application from Existing Maven Binaries and Securing it Using Red Hat Single Sing-On

== Tutorials

[[Example-Deploying-SSO]]
=== Example Workflow: Preparing and Deploying the {project_openshift_product_name} image
[[Preparing-SSO-Authentication-for-OpenShift-Deployment]]
==== Preparing {project_name} Authentication for OpenShift Deployment
Log in to the OpenShift CLI with a user that holds the _cluster:admin_ role.

To deploy existing applications on OpenShift, you can use the link:https://docs.openshift.com/container-platform/latest/dev_guide/builds/build_inputs.html#binary-source[binary source] capability.

==== Deploy Binary Build of EAP 6.4 / 7.1 JSP Service Invocation Application and Secure it Using Red Hat Single Sign-On

The following example uses both link:https://github.com/keycloak/keycloak-quickstarts/tree/latest/app-jee-jsp[app-jee-jsp] and link:https://github.com/keycloak/keycloak-quickstarts/tree/latest/service-jee-jaxrs[service-jee-jaxrs] quickstarts to deploy EAP 6.4 / 7.1 JSP service application that authenticates using the Red Hat Single Sign-On.

*Prerequisite:*

[IMPORTANT]
====
This guide assumes the {project_openshift_product_name} image has been previously link:https://access.redhat.com/documentation/en-us/red_hat_jboss_middleware_for_openshift/3/html-single/red_hat_single_sign-on_for_openshift/index#Example-Deploying-SSO[deployed using one of the following templates:]

* *_{project_templates_version}-mysql_*
* *_{project_templates_version}-mysql-persistent_*
* *_{project_templates_version}-postgresql_*
* *_{project_templates_version}-postgresql-persistent_*
* *_{project_templates_version}-x509-mysql-persistent_*
* *_{project_templates_version}-x509-postgresql-persistent_*
====

===== Create {project_name} Realm, Roles, and User for the EAP 6.4 / 7.1 JSP Application

The EAP 6.4 / 7.1 JSP service application requires dedicated {project_name} realm, username, and password to be able to authenticate using Red Hat Single Sign-On. Perform the following steps after the {project_openshift_product_name} image has been deployed:

*Create the {project_name} Realm*

. Login to the administration console of the {project_name} server.
+
*\https://secure-sso-sso-app-demo.openshift.example.com/auth/admin*
+
Use the xref:sso-administrator-setup[credentials of the {project_name} administrator user].
. Hover your cursor over the realm namespace (default is *Master*) at the top of the sidebar and click *Add Realm*.
. Enter a realm name (this example uses `demo`) and click *Create*.


[[upgrading-sso-db-from-previous-to-{project_version}]]
=== Example Workflow: Updating Existing Database when Migrating {project_openshift_product_name} Image to a new version
[IMPORTANT]
====
* Rolling updates from previous versions of {project_openshift_product_name} to version {project_version} are not supported as databases and caches are not backward compatible.
* Stop all instances running some of previous versions of the {project_openshift_product_name} before upgrading. They cannot run concurrently against the same database.
* Pre-generated scripts are not available, they are generated dynamically depending on the database.
====

{project_name} {project_version} can xref:automatic-db-migration[automatically migrate the database schema], or you can choose to do it xref:manual-db-migration[manually].

[NOTE]
====
By default the database is automatically migrated when you start {project_name} {project_version} for the first time.
====

[[automatic-db-migration]]
==== Automatic Database Migration
This process assumes that you are link:https://access.redhat.com/documentation/en-us/red_hat_jboss_middleware_for_openshift/3/html-single/red_hat_single_sign-on_for_openshift/index#Example-Deploying-SSO[running] some previous version of the {project_openshift_product_name} image, backed by MySQL or PostgreSQL database (deployed in ephemeral or persistent mode), running on a separate pod.

[IMPORTANT]
====
Stop all pods running the previous version of the {project_openshift_product_name} image before upgrading to {project_name} {project_version}, as they cannot run concurrently against the same database.
====

Use the following steps to automatically migrate the database schema:

. Identify the deployment config used to deploy the containers, running previous version of the {project_openshift_product_name} image.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get dc -o name --selector=application=sso
deploymentconfig/sso
deploymentconfig/sso-postgresql
----
. Stop all pods running the previous version of the {project_openshift_product_name} image in the current namespace.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=0 dc/sso
deploymentconfig "sso" scaled
----
. Update the image change trigger in the existing deployment config to reference the {project_name} {project_version} image.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc patch dc/sso --type=json -p '[{"op": "replace", "path": "/spec/triggers/0/imageChangeParams/from/name", "value": "redhat-{project_templates_version}-openshift:{project_latest_image_tag}"}]'
"sso" patched
----
. Start rollout of the new {project_name} {project_version} images based on the latest image defined in the image change triggers.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rollout latest dc/sso
deploymentconfig "sso" rolled out
----
. Deploy {project_name} {project_version} containers using the modified deployment config.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=1 dc/sso
deploymentconfig "sso" scaled
----
. (Optional) Verify the database has been successfully updated.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get pods --selector=application=sso
NAME                     READY     STATUS    RESTARTS   AGE
sso-4-vg21r              1/1       Running   0          1h
sso-postgresql-1-t871r   1/1       Running   0          2h
----
+
[source,bash,subs="attributes+,macros+"]
----
$ oc logs sso-4-vg21r | grep 'Updating'
11:23:45,160 INFO  [org.keycloak.connections.jpa.updater.liquibase.LiquibaseJpaUpdaterProvider] (ServerService Thread Pool -- 58) Updating database. Using changelog META-INF/jpa-changelog-master.xml
----

[[manual-db-migration]]
==== Manual Database Migration

[IMPORTANT]
====
Pre-generated scripts are not available. They are generated dynamically depending on the database. With {project_name} {project_version} one can generate and export these to an SQL file that can be manually applied to the database afterwards. To dynamically generate the SQL migration file for the database:

. Configure {project_name} {project_version} with the correct datasource,
. Set following configuration options in the `standalone-openshift.xml` file:
.. `initializeEmpty=false`,
.. `migrationStrategy=manual`, and
.. `migrationExport` to the location on the file system of the pod, where the output SQL migration file should be stored (e.g. `migrationExport="${jboss.home.dir}/keycloak-database-update.sql"`).

See link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html/server_installation_and_configuration_guide/database-1#database_configuration[database configuration of {project_name} {project_version}] for further details.

The database migration process handles the data schema update and performs manipulation of the data, therefore, stop all pods running the previous version of the {project_openshift_product_name} image before dynamic generation of the SQL migration file.
====

This process assumes that you are link:https://access.redhat.com/documentation/en-us/red_hat_jboss_middleware_for_openshift/3/html-single/red_hat_single_sign-on_for_openshift/index#Example-Deploying-SSO[running] some previous version of the {project_openshift_product_name} image, backed by MySQL or PostgreSQL database (deployed in ephemeral or persistent mode), running on a separate pod.

Perform the following to generate and get the SQL migration file for the database:

. Prepare template of OpenShift link:https://docs.openshift.com/container-platform/latest/dev_guide/jobs.html[database migration job] to generate the SQL file.
+
[source,yaml,subs="verbatim,macros,attributes"]
----
$ cat job-to-migrate-db-to-{project_templates_version}.yaml.orig
apiVersion: batch/v1
kind: Job
metadata:
  name: job-to-migrate-db-to-{project_templates_version}
spec:
  autoSelector: true
  parallelism: 0
  completions: 1
  template:
    metadata:
      name: job-to-migrate-db-to-{project_templates_version}
    spec:
      containers:
      - env:
        - name: DB_SERVICE_PREFIX_MAPPING
          value: pass:[&lt;&lt;DB_SERVICE_PREFIX_MAPPING_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;PREFIX&gt;&gt;]_JNDI
          value: pass:[&lt;&lt;PREFIX_JNDI_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;PREFIX&gt;&gt;]_USERNAME
          value: pass:[&lt;&lt;PREFIX_USERNAME_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;PREFIX&gt;&gt;]_PASSWORD
          value: pass:[&lt;&lt;PREFIX_PASSWORD_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;PREFIX&gt;&gt;]_DATABASE
          value: pass:[&lt;&lt;PREFIX_DATABASE_VALUE&gt;&gt;]
        - name: TX_DATABASE_PREFIX_MAPPING
          value: pass:[&lt;&lt;TX_DATABASE_PREFIX_MAPPING_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;SERVICE_HOST&gt;&gt;]
          value: pass:[&lt;&lt;SERVICE_HOST_VALUE&gt;&gt;]
        - name: pass:[&lt;&lt;SERVICE_PORT&gt;&gt;]
          value: pass:[&lt;&lt;SERVICE_PORT_VALUE&gt;&gt;]
        image: pass:[&lt;&lt;SSO_IMAGE_VALUE&gt;&gt;]
        imagePullPolicy: Always
        name: job-to-migrate-db-to-{project_templates_version}
        # Keep the pod running after the SQL migration
        # file was generated, so we can retrieve it
        command:
          - "/bin/bash"
          - "-c"
          - "/opt/eap/bin/openshift-launch.sh || sleep 600"
      restartPolicy: Never
----
+
[source,bash,subs="attributes+,macros+"]
----
$ cp job-to-migrate-db-to-{project_templates_version}.yaml.orig \
     job-to-migrate-db-to-{project_templates_version}.yaml
----
. From deployment config used to run the previous version of the {project_openshift_product_name} image, copy the datasource definition and database access credentials to appropriate places of the template of the database migration job.
+
Use the following script to copy `DB_SERVICE_PREFIX_MAPPING` and `TX_DATABASE_PREFIX_MAPPING` variable values, together with values of environment variables specific to particular datasource (`<PREFIX>_JNDI`, `<PREFIX>_USERNAME`, `<PREFIX>_PASSWORD`, and `<PREFIX>_DATABASE`) from the deployment config named `sso` to the database job migration template named `job-to-migrate-db-to-{project_templates_version}.yaml`.
+
[NOTE]
====
Although the `DB_SERVICE_PREFIX_MAPPING` environment variable allows a link:https://access.redhat.com/documentation/en-us/red_hat_jboss_enterprise_application_platform/7.1/html-single/red_hat_jboss_enterprise_application_platform_for_openshift/#datasources[comma-separated list of *<name>-<database_type>=<PREFIX>* triplets] as its value, this example script accepts only one datasource triplet definition for demonstration purposes. You can modify the script for handling multiple datasource definition triplets.
====
+
[source,bash,subs="verbatim,macros,attributes"]
----
$ cat mirror_sso_dc_db_vars.sh
#!/bin/bash

# IMPORTANT:
#
# If the name of the SSO deployment config differs from 'sso'
# or if the file name of the YAML definition of the migration
# job is different, update the following two variables
SSO_DC_NAME="sso"
JOB_MIGRATION_YAML="job-to-migrate-db-to-{project_templates_version}.yaml"

# Get existing variables of the $SSO_DC_NAME deployment config
# in an array
declare -a SSO_DC_VARS=( \
  $(oc set env dc/${SSO_DC_NAME} --list \
  | sed '/^#/d') \
)

# Get the PREFIX used in the names of environment variables
PREFIX=$( \
  grep -oP 'DB_SERVICE_PREFIX_MAPPING=\[^ ]++' \
  <<< "${SSO_DC_VARS[@]}" \
)
PREFIX=${PREFIX##*=}

# Substitute:
# * <<PREFIX>> with actual $PREFIX value and
# * <<PREFIX with "<<$PREFIX" value
# The order in which these replacements are made is important!
sed -i "s#<<PREFIX>>#${PREFIX}#g" ${JOB_MIGRATION_YAML}
sed -i "s#<<PREFIX#<<${PREFIX}#g" ${JOB_MIGRATION_YAML}

# Construct the array of environment variables
# specific to the datasource
declare -a DB_VARS=(JNDI USERNAME PASSWORD DATABASE)

# Prepend $PREFIX to each item of the datasource array
DB_VARS=( "${DB_VARS[@]/#/${PREFIX}_}" )

# Add DB_SERVICE_PREFIX_MAPPING and TX_DATABASE_PREFIX_MAPPING
# variables to datasource array
DB_VARS=( \
  "${DB_VARS[@]}" \
  DB_SERVICE_PREFIX_MAPPING \
  TX_DATABASE_PREFIX_MAPPING \
)

# Construct the SERVICE from DB_SERVICE_PREFIX_MAPPING
SERVICE=$( \
  grep -oP 'DB_SERVICE_PREFIX_MAPPING=[^ ]+' \
  <<< "${SSO_DC_VARS[@]}" \
)
SERVICE=${SERVICE#*=}
SERVICE=${SERVICE%=*}
SERVICE=${SERVICE^^}
SERVICE=${SERVICE//-/_}

# If the deployment config contains pass:[&lt;&lt;SERVICE&gt;&gt;]_SERVICE_HOST
# and pass:[&lt;&lt;SERVICE&gt;&gt;]_SERVICE_PORT variables, add them to the
# datasource array. Their values also need to be propagated into
# yaml definition of the migration job.
HOST_PATTERN="${SERVICE}_SERVICE_HOST=\[^ ]+"
PORT_PATTERN="${SERVICE}_SERVICE_PORT=[^ ]+"
if
  grep -Pq "${HOST_PATTERN}" <<< "${SSO_DC_VARS[@]}" &&
  grep -Pq "${PORT_PATTERN}" <<< "${SSO_DC_VARS[@]}"
then
  DB_VARS=( \
    "${DB_VARS[@]}" \
    "${SERVICE}_SERVICE_HOST" \
    "${SERVICE}_SERVICE_PORT" \
  )
# If they are not defined, delete their placeholder rows in
# yaml definition file (since if not defined they are not
# expanded which make the yaml definition invalid).
else
  for KEY in "HOST" "PORT"
  do
    sed -i "/SERVICE_${KEY}/d" ${JOB_MIGRATION_YAML}
  done
fi

# Substitute:
# * pass:[&lt;&lt;SERVICE_HOST&gt;&gt;] with ${SERVICE}_SERVICE_HOST and
# * pass:[&lt;&lt;SERVICE_HOST_VALUE&gt;&gt;] with pass:[&lt;&lt;${SERVICE}_SERVICE_HOST_VALUE&gt;&gt;]
# The order in which replacements are made is important!
# Do this for both "HOST" and "PORT"
for KEY in "HOST" "PORT"
do
  PATTERN_1=pass:["&lt;&lt;SERVICE_${KEY}&gt;&gt;"]
  REPL_1="${SERVICE}_SERVICE_${KEY}"
  sed -i "s#${PATTERN_1}#${REPL_1}#g" ${JOB_MIGRATION_YAML}
  PATTERN_2=pass:["&lt;&lt;SERVICE_${KEY}_VALUE&gt;&gt;"]
  REPL_2="<<${SERVICE}_SERVICE_${KEY}_VALUE>>"
  sed -i "s#${PATTERN_2}#${REPL_2}#g" ${JOB_MIGRATION_YAML}
done

# Propagate the values of the datasource array items into
# yaml definition of the migration job
for VAR in "${SSO_DC_VARS[@]}"
do
  IFS=$'=' read KEY VALUE <<< $VAR
  if grep -q $KEY <<< ${DB_VARS[@]}
  then
    KEY+="_VALUE"
    # Enwrap integer port value with double quotes
    if [[ ${KEY} =~ ${SERVICE}_SERVICE_PORT_VALUE ]]
    then
      sed -i "s#<<${KEY}>>#\"${VALUE}\"#g" ${JOB_MIGRATION_YAML}
    # Character values do not need quotes
    else
      sed -i "s#<<${KEY}>>#${VALUE}#g" ${JOB_MIGRATION_YAML}
    fi
    # Verify that the value has been successfully propagated.
    if
      grep -q '(JNDI|USERNAME|PASSWORD|DATABASE)' <<< "${KEY}" &&
      pass:[grep -q "&lt;&lt;PREFIX${KEY#${PREFIX}}"] ${JOB_MIGRATION_YAML} ||
      grep -q "<<${KEY}>>" ${JOB_MIGRATION_YAML}
    then
      echo "Failed to update value of ${KEY%_VALUE}! Aborting."
      exit 1
    else
      printf '%-60s%-40s\n' \
             "Successfully updated ${KEY%_VALUE} to:" \
             "$VALUE"
    fi
  fi
done
----
+
[[get-db-credentials]]
Run the script.
+
[source,bash,subs="attributes+,macros+"]
----
$ chmod +x ./mirror_sso_dc_db_vars.sh
$ ./mirror_sso_dc_db_vars.sh
Successfully updated DB_SERVICE_PREFIX_MAPPING to:          sso-postgresql=DB
Successfully updated DB_JNDI to:                            java:jboss/datasources/KeycloakDS
Successfully updated DB_USERNAME to:                        userxOp
Successfully updated DB_PASSWORD to:                        tsWNhQHK
Successfully updated DB_DATABASE to:                        root
Successfully updated TX_DATABASE_PREFIX_MAPPING to:         sso-postgresql=DB
----
. Build the {project_name} {project_version} database migration image using the link:https://github.com/iankko/openshift-examples/tree/KEYCLOAK-8500/sso-manual-db-migration[pre-configured source] and wait for the build to finish.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get is -n openshift | grep {project_templates_version} | cut -d ' ' -f1
redhat-{project_templates_version}-openshift
----
+
[source,bash,subs="attributes+,macros+"]
----
$ oc new-build redhat-{project_templates_version}-openshift:{project_latest_image_tag}~https://github.com/iankko/openshift-examples.git#KEYCLOAK-8500 \
  --context-dir=sso-manual-db-migration \
  --name={project_templates_version}-db-migration-image
--> Found image bf45ac2 (7 days old) in image stream "openshift/redhat-{project_templates_version}-openshift" under tag "{project_latest_image_tag}" for "redhat-{project_templates_version}-openshift:{project_latest_image_tag}"

    Red Hat SSO {project_version}
    ---------------
    Platform for running Red Hat SSO

    Tags: sso, sso7, keycloak

    * A source build using source code from \https://github.com/iankko/openshift-examples.git#KEYCLOAK-8500 will be created
      * The resulting image will be pushed to image stream "{project_templates_version}-db-migration-image:latest"
      * Use 'start-build' to trigger a new build

--> Creating resources with label build={project_templates_version}-db-migration-image ...
    imagestream "{project_templates_version}-db-migration-image" created
    buildconfig "{project_templates_version}-db-migration-image" created
--> Success
    Build configuration "{project_templates_version}-db-migration-image" created and build triggered.
    Run 'oc logs -f bc/{project_templates_version}-db-migration-image' to stream the build progress.
----
+
[source,bash,subs="attributes+,macros+"]
----
$ oc logs -f bc/{project_templates_version}-db-migration-image --follow
Cloning "https://github.com/iankko/openshift-examples.git#KEYCLOAK-8500" ...
...
Push successful
----
. Update the template of the database migration job (`job-to-migrate-db-to-{project_templates_version}.yaml`) with reference to the built `{project_templates_version}-db-migration-image` image.
.. Get the docker pull reference for the image.
+
[source,bash,subs="attributes+,macros+"]
----
$ PULL_REF=$(oc get istag -n $(oc project -q) --no-headers | grep {project_templates_version}-db-migration-image | tr -s ' ' | cut -d ' ' -f 2)
----
.. Replace the pass:[&lt;&lt;SSO_IMAGE_VALUE&gt;&gt;] field in the job template with the pull specification.
+
[source,bash,subs="attributes+,macros+"]
----
$ sed -i "s#pass:[&lt;&lt;SSO_IMAGE_VALUE&gt;&gt;]#$PULL_REF#g" job-to-migrate-db-to-{project_templates_version}.yaml
----
.. Verify that the field is updated.
. Instantiate database migration job from the job template.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc create -f job-to-migrate-db-to-{project_templates_version}.yaml
job "job-to-migrate-db-to-{project_templates_version}" created
----
+
[IMPORTANT]
====
The database migration process handles the data schema update and performs manipulation of the data, therefore, stop all pods running the previous version of the {project_openshift_product_name} image before dynamic generation of the SQL migration file.
====
+
. Identify the deployment config used to deploy the containers, running previous version of the {project_openshift_product_name} image.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get dc -o name --selector=application=sso
deploymentconfig/sso
deploymentconfig/sso-postgresql
----
. Stop all pods running the previous version of the {project_openshift_product_name} image in the current namespace.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=0 dc/sso
deploymentconfig "sso" scaled
----
. Run the database migration job and wait for the pod to be running correctly.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get jobs
NAME                            DESIRED   SUCCESSFUL   AGE
job-to-migrate-db-to-{project_templates_version}   1         0            3m
----
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=1 job/job-to-migrate-db-to-{project_templates_version}
job "job-to-migrate-db-to-{project_templates_version}" scaled
----
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get pods
NAME                                  READY     STATUS      RESTARTS   AGE
sso-postgresql-1-n5p16                1/1       Running     1          19h
job-to-migrate-db-to-{project_templates_version}-b87bb   1/1       Running     0          1m
{project_templates_version}-db-migration-image-1-build      0/1       Completed   0          27m
----
+
[NOTE]
====
By default, the database migration job terminates automatically after `600 seconds` after the migration file is generated. You can adjust this time period.
====
. Get the dynamically generated SQL database migration file from the pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ mkdir -p ./db-update
$ oc rsync job-to-migrate-db-to-{project_templates_version}-b87bb:/opt/eap/keycloak-database-update.sql ./db-update
receiving incremental file list
keycloak-database-update.sql

sent 30 bytes  received 29,726 bytes  59,512.00 bytes/sec
total size is 29,621  speedup is 1.00
----
. Inspect the `keycloak-database-update.sql` file for changes to be performed within manual database update to {project_name} {project_version} version.
. Apply the database update manually.
* Run the following commands if running some previous version of the {project_openshift_product_name} image, backed by the PostgreSQL database deployed in ephemeral or persistent mode, running on a separate pod:
... Copy the generated SQL migration file to the PostgreSQL pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rsync --no-perms=true ./db-update/ sso-postgresql-1-n5p16:/tmp
sending incremental file list

sent 77 bytes  received 11 bytes  176.00 bytes/sec
total size is 26,333  speedup is 299.24
----
... Start a shell session to the PostgreSQL pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rsh sso-postgresql-1-n5p16
sh-4.2$
----
... Use the `psql` tool to apply database update manually.
+
[source,bash,subs="attributes+,macros+"]
----
sh-4.2$ alias psql="/opt/rh/rh-postgresql95/root/bin/psql"
sh-4.2$ psql --version
psql (PostgreSQL) 9.5.4
sh-4.2$ psql -U <PREFIX>_USERNAME -d <PREFIX>_DATABASE -W -f /tmp/keycloak-database-update.sql
Password for user <PREFIX>_USERNAME:
INSERT 0 1
INSERT 0 1
...
----
+
[IMPORTANT]
====
Replace `<PREFIX>_USERNAME` and `<PREFIX>_DATABASE` with the actual database credentials retrieved xref:get-db-credentials[in previous section]. Also use value of `<PREFIX>_PASSWORD` as the password for the database, when prompted.
====
... Close the shell session to the PostgreSQL pod. Continue with xref:image-change-trigger-update-step[updating image change trigger step].
* Run the following commands if running some previous version of the {project_openshift_product_name} image, backed by the MySQL database deployed in ephemeral or persistent mode, running on a separate pod:
... Given the pod situation similar to the following:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get pods
NAME                                  READY     STATUS      RESTARTS   AGE
sso-mysql-1-zvhk3                     1/1       Running     0          1h
job-to-migrate-db-to-{project_templates_version}-m202t   1/1       Running     0          11m
{project_templates_version}-db-migration-image-1-build      0/1       Completed   0          13m
----
... Copy the generated SQL migration file to the MySQL pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rsync --no-perms=true ./db-update/ sso-mysql-1-zvhk3:/tmp
sending incremental file list
keycloak-database-update.sql

sent 24,718 bytes  received 34 bytes  49,504.00 bytes/sec
total size is 24,594  speedup is 0.99
----
... Start a shell session to the MySQL pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rsh sso-mysql-1-zvhk3
sh-4.2$
----
... Use the `mysql` tool to apply database update manually.
+
[source,bash,subs="attributes+,macros+"]
----
sh-4.2$ alias mysql="/opt/rh/rh-mysql57/root/bin/mysql"
sh-4.2$ mysql --version
/opt/rh/rh-mysql57/root/bin/mysql  Ver 14.14 Distrib 5.7.16, for Linux (x86_64) using  EditLine wrapper
sh-4.2$ mysql -D <PREFIX>_DATABASE -u <PREFIX>_USERNAME -p < /tmp/keycloak-database-update.sql
Enter password:
sh-4.2$ echo $?
0
----
+
[IMPORTANT]
====
Replace `<PREFIX>_USERNAME` and `<PREFIX>_DATABASE` with the actual database credentials retrieved xref:get-db-credentials[in previous section]. Also use value of `<PREFIX>_PASSWORD` as the password for the database, when prompted.
====
... Close the shell session to the MySQL pod. Continue with xref:image-change-trigger-update-step[updating image change trigger step].

[[image-change-trigger-update-step]]
[start=12]
. Update the image change trigger in the existing deployment config to reference the {project_name} {project_version} image.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc patch dc/sso --type=json -p '[{"op": "replace", "path": "/spec/triggers/0/imageChangeParams/from/name", "value": "redhat-{project_templates_version}-openshift:{project_latest_image_tag}"}]'
"sso" patched
----
. Start rollout of the new {project_name} {project_version} images based on the latest image defined in the image change triggers.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc rollout latest dc/sso
deploymentconfig "sso" rolled out
----
. Deploy the {project_name} {project_version} containers using the modified deployment config.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=1 dc/sso
deploymentconfig "sso" scaled
----

=== Example Workflow: Migrating Entire {project_name} Server Database Across The Environments
This tutorial focuses on migrating the Red Hat Single Sign-On server database from one environment to another or migrating to a different database. It assumes steps described in xref:Preparing-SSO-Authentication-for-OpenShift-Deployment[Preparing {project_name} Authentication for OpenShift Deployment] section have been performed already.

==== Deploying the {project_name} MySQL Application Template

. Log in to the OpenShift web console and select the _sso-app-demo_ project space.
. Click *Add to project* to list the default image streams and templates.
. Use the *Filter by keyword* search bar to limit the list to those that match _sso_. You may need to click *See all* to show the desired application template.
. Select *_{project_templates_version}-mysql_* {project_name} application template. When deploying the template ensure to *keep the _SSO_REALM_ variable unset* (default value).
+
[IMPORTANT]
====
Export and import of {project_name} {project_version} database link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#export_import[is triggered at {project_name} server boot time and its paramaters are passed in via Java system properties.] This means during one {project_name} server boot only one of the possible migration actions (either *_export_*, or *_import_*) can be performed.
====
+
[WARNING]
====
When the *_SSO_REALM_* configuration variable is set on the {project_openshift_product_name} image, a database import is performed in order to create the {project_name} server realm requested in the variable. For the database export to be performed correctly, the *_SSO_REALM_* configuration variable cannot be simultaneously defined on such image.
====
+
. Click *Create* to deploy the application template and start pod deployment. This may take a couple of minutes.
+
Then access the {project_name} web console at *$$https://secure-sso-$$_<sso-app-demo>_._<openshift32.example.com>_/auth/admin* using the xref:sso-administrator-setup[administrator account].
+
[NOTE]
====
This example workflow uses a self-generated CA to provide an end-to-end workflow for demonstration purposes. Accessing the {project_name} web console will prompt an insecure connection warning. +
For production environments, Red Hat recommends that you use an SSL certificate purchased from a verified Certificate Authority.
====

==== (Optional) Creating additional {project_name} link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#core_concepts_and_terms[realm and users] to be also exported

When performing link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#export_import[{project_name} {project_version} server database export] only {project_name} realms and users currently present in the database will be exported. If the exported JSON file should include also additional {project_name} realms and users, these need to be created first:

. link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#create-realm[Create a new realm]
. link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#create-new-user[Create new users]

Upon their creation xref:sso-export-the-database[the database can be exported.]

[[sso-export-the-database]]
==== Export the {project_name} database as a JSON file on the OpenShift pod

. Get the {project_name} deployment config and scale it down to zero.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get dc -o name
deploymentconfig/sso
deploymentconfig/sso-mysql

$ oc scale --replicas=0 dc sso
deploymentconfig "sso" scaled
----
. Instruct the {project_name} {project_version} server deployed on {project_openshift_product_name} image to perform database export at {project_name} server boot time.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc set env dc/sso \
  -e "JAVA_OPTS_APPEND= \
    -Dkeycloak.migration.action=export \
    -Dkeycloak.migration.provider=singleFile \
    -Dkeycloak.migration.file=/tmp/demorealm-export.json"
----
. Scale the {project_name} deployment config back up. This will start the {project_name} server and export its database.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc scale --replicas=1 dc sso
deploymentconfig "sso" scaled
----
. (Optional) Verify that the export was successful.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get pods
NAME                READY     STATUS    RESTARTS   AGE
sso-4-ejr0k         1/1       Running   0          27m
sso-mysql-1-ozzl0   1/1       Running   0          4h

$ oc logs sso-4-ejr0k | grep 'Export'
09:24:59,503 INFO  [org.keycloak.exportimport.singlefile.SingleFileExportProvider] (ServerService Thread Pool -- 57) Exporting model into file /tmp/demorealm-export.json
09:24:59,998 INFO  [org.keycloak.services] (ServerService Thread Pool -- 57) KC-SERVICES0035: Export finished successfully
----

==== Retrieve and import the exported JSON file

. Retrieve the JSON file of the {project_name} database from the pod.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc get pods
NAME                READY     STATUS    RESTARTS   AGE
sso-4-ejr0k         1/1       Running   0          2m
sso-mysql-1-ozzl0   1/1       Running   0          4h

$ oc rsync sso-4-ejr0k:/tmp/demorealm-export.json .
----

. (Optional) Import the JSON file of the {project_name} database into an {project_name} server running in another environment.
+
[NOTE]
====
For importing into an {project_name} server not running on OpenShift, see the link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html/server_administration_guide/export_import[Export and Import section] of the RH SSO Server Administration Guide.
====
+
Use the link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#admin_console_export_import[administration console] of the {project_name} server to import the resources from previously exported JSON file into the {project_name} server's database, when the {project_name} server is running as a {project_name} {project_version} container on OpenShift:

.. Log into the `master` realm's administration console of the {project_name} server using the credentials used to create the administrator user. In the browser, navigate to *\http://sso-<project-name>.<hostname>/auth/admin*  for the {project_name} web server, or to *\https://secure-sso-<project-name>.<hostname>/auth/admin* for the encrypted {project_name} web server.
.. At the top of the sidebar choose the name of the {project_name} realm, the users, clients, realm roles, and client roles should be imported to. This example uses `master` realm.
.. Click the *Import* link under *Manage* section at the bottom of the sidebar.
.. In the page that opens, click *Select file* and then specify the location of the exported `demorealm-export.json` JSON file on the local file system.
.. From the *Import from realm* drop-down menu, select the name of the {project_name} realm from which the data should be imported. This example uses `master` realm.
.. Choose which of users, clients, realm roles, and client roles should be imported (all of them are imported by default).
.. Choose a strategy to perform, when a resource already exists (one of *Fail*, *Skip*, or *Overwrite*).
+
[NOTE]
====
The attempt to import an object (user, client, realm role, or client role) fails if object with the same identifier already exists in the current database. Use *Skip* strategy to import the objects that are present in the `demorealm-export.json` file, but do not exist in current database.
====
.. Click *Import* to perform the import.
+
[NOTE]
====
When importing objects from a non-master realm to `master` realm or vice versa, after clicking the *Import* button, it is sometimes possible to encounter an error like the following one:

[[realm-import-error-message]]
[.text-center]
image:images/import_realm_error.png[Example of Possible Error Message when Performing Partial Import from Previously Exported JSON File]

In such cases, it is necessary first to create the missing clients, having the *Access Type* set to *bearer-only*. These clients can be created by manual copy of their characteristics from the source {project_name} server, on which the export JSON file was created, to the target {project_name} server, where the JSON file is imported. After creation of the necessary clients, click the *Import* button again.

To suppress the xref:realm-import-error-message[above] error message, it is needed to create the missing `realm-management` client, of the *bearer-only* *Access Type*, and click the *Import* button again.
====
+
[NOTE]
====
For *Skip* import strategy, the newly added objects are marked as *ADDED* and the object which were skipped are marked as *SKIPPED*, in the *Action* column on the import result page.
====
+
[IMPORTANT]
====
The administration console import allows you to *overwrite* resources if you choose (*Overwrite* strategy). On a production system use this feature with caution.
====

[[OSE-SSO-AUTH-TUTE]]
=== Example Workflow: Configuring OpenShift to use {project_name} for Authentication
Configure OpenShift to use the {project_name} deployment as the authorization gateway for OpenShift. This follows on from xref:Example-Deploying-SSO[Example Workflow: Preparing and Deploying the {project_openshift_product_name} image], in which {project_name} was deployed on OpenShift.

This example adds {project_name} as an authentication method alongside https://docs.openshift.com/container-platform/latest/install_config/configuring_authentication.html#identity-providers-configuring[the identity providers] configured during https://docs.openshift.com/container-platform/latest/install/index.html[the installation of the OpenShift Container Platform cluster]. Once configured, the {project_name} method will be also available (together with the configured identity providers) for the user login to your OpenShift web console.

==== Configuring {project_name} Credentials
Log in to the encrypted {project_name} web server at *$$https://secure-sso-$$_sso-app-demo_._openshift32.example.com_/auth/admin* using the xref:sso-administrator-setup[administrator account] created during the {project_name} deployment.

*Create a Realm*

. Hover your cursor over the realm namespace (default is *Master*) at the top of the sidebar and click *Add Realm*.
. Enter a realm name (this example uses _OpenShift_) and click *Create*.

*Create a User*

Create a test user that can be used to demonstrate the {project_name}-enabled OpenShift login:

. Click *Users* in the *Manage* sidebar to view the user information for the realm.
. Click *Add User*.
. Enter a valid *Username* (this example uses _testuser_) and any additional optional information and click *Save*.
. Edit the user configuration:
.. Click the *Credentials* tab in the user space and enter a password for the user.
.. Ensure the *Temporary Password* option is set to *Off* so that it does not prompt for a password change later on, and click *Reset Password* to set the user password. A pop-up window prompts for additional confirmation.

*Create and Configure an OpenID-Connect Client*

See the link:https://access.redhat.com/documentation/en-us/red_hat_single_sign-on/7.2/html-single/server_administration_guide/#clients[Managing Clients] chapter of the Red Hat Single Sign-On Server Administration Guide for more information.

. Click *Clients* in the *Manage* sidebar and click *Create*.
. Enter the *Client ID*. This example uses _openshift-demo_.
. Select a *Client Protocol* from the drop-down menu (this example uses *openid-connect*) and click *Save*. You will be taken to the configuration *Settings* page of the _openshift-demo_ client.
. From the *Access Type* drop-down menu, select *confidential*. This is the access type for server-side applications.
. In the *Valid Redirect URIs* dialog, enter the URI for the OpenShift web console, which is _$$https://openshift$$.example.com:8443/*_ in this example.

The client *Secret* is needed to configure OpenID-Connect on the OpenShift master in the next section. You can copy it now from under the *Credentials* tab. The secret is <pass:quotes[_7b0384a2-b832-16c5-9d73-2957842e89h7_]> for this example.

==== Configuring OpenShift Master for {project_name} Authentication
Log in to the OpenShift master CLI. You must have the required permissions to edit the */etc/origin/master/master-config.yaml* file.

. Edit the */etc/origin/master/master-config.yaml* file and find the *identityProviders*. For example, in the case the OpenShift master is configured with the https://docs.openshift.com/container-platform/latest/install_config/configuring_authentication.html#HTPasswdPasswordIdentityProvider[HTPassword identity provider], the *identityProviders* section will look similar to the following one:
+
[source,bash,subs="attributes+,macros+"]
----
identityProviders:
- challenge: true
  login: true
  name: htpasswd_auth
  provider:
    apiVersion: v1
    file: /etc/origin/openshift-passwd
    kind: HTPasswdPasswordIdentityProvider
----
+
Add {project_name} as a secondary identity provider with content similar to the following snippet:
+
[source,bash,subs="attributes+,macros+"]
----
- name: rh_sso
  challenge: false
  login: true
  mappingMethod: add
  provider:
    apiVersion: v1
    kind: OpenIDIdentityProvider
    clientID: pass:quotes[_openshift-demo_]
    clientSecret: pass:quotes[_7b0384a2-b832-16c5-9d73-2957842e89h7_]
    pass:quotes[_ca: xpaas.crt_]
    urls:
      authorize: pass:quotes[_https://secure-sso-sso-app-demo.openshift32.example.com/auth/realms/OpenShift/protocol/openid-connect/auth_]
      token: pass:quotes[_https://secure-sso-sso-app-demo.openshift32.example.com/auth/realms/OpenShift/protocol/openid-connect/token_]
      userInfo: pass:quotes[_https://secure-sso-sso-app-demo.openshift32.example.com/auth/realms/OpenShift/protocol/openid-connect/userinfo_]
    claims:
      id:
      - sub
      preferredUsername:
      - preferred_username
      name:
      - name
      email:
      - email
----
.. The {project_name} *Secret* hash for the *clientSecret* can be found in the {project_name} web console: *Clients* -> *_openshift-demo_* -> *Credentials*
.. The endpoints for the *urls* can be found by making a request with the {project_name} application. For example:
+
[source,bash,subs="attributes+,macros+"]
----
<pass:quotes[_curl -k https://secure-sso-sso-app-demo.openshift32.example.com/auth/realms/OpenShift/.well-known/openid-configuration | python -m json.tool_]>
----
+
The response includes the *authorization_endpoint*, *token_endpoint*, and *userinfo_endpoint*.
+
.. This example workflow uses a self-generated CA to provide an end-to-end workflow for demonstration purposes. For this reason, the *ca* is provided as <pass:quotes[_ca: xpaas.crt_]>. This CA certificate must also be copied into the */etc/origin/master* folder. This is not necessary if using a certificate purchased from a verified Certificate Authority.
. Save the configuration and restart the OpenShift master:
+
[source,bash,subs="attributes+,macros+"]
----
$ systemctl restart atomic-openshift-master
----

==== Logging in to OpenShift

Navigate to the OpenShift web console, which in this example is _https://openshift.example.com:8443/console_. The OpenShift login page now has the option to use either *htpasswd_auth* or *rh-sso*. The former is still available because it is present in the */etc/origin/master/master-config.yaml*.

Select *rh-sso* and log in to OpenShift with the _testuser_ user created earlier in {project_name}. No projects are visible to _testuser_ until they are added in the OpenShift CLI. This is the only way to provide user privileges in OpenShift because it currently does not accept external role mapping.

To provide _testuser_ `view` privileges for the _sso-app-demo_, use the OpenShift CLI:
[source,bash,subs="attributes+,macros+"]
----
$ oc adm policy add-role-to-user view testuser -n sso-app-demo
----

[[Example-EAP-Auto]]
=== Example Workflow: Automatically Registering EAP Application in {project_name} with OpenID-Connect Client
This follows on from xref:Example-Deploying-SSO[Example Workflow: Preparing and Deploying the {project_openshift_product_name} image], in which {project_name} was deployed on OpenShift. This example prepares {project_name} realm, role, and user credentials for an EAP project using an OpenID-Connect client adapter. These credentials are then provided in the EAP for OpenShift template for automatic {project_name} client registration. Once deployed, the {project_name} user can be used to authenticate and access JBoss EAP.

[NOTE]
====
This example uses a OpenID-Connect client but an SAML client could also be used. See xref:../advanced_concepts/advanced_concepts.adoc#SSO-Clients[{project_name} Clients] and xref:../advanced_concepts/advanced_concepts.adoc#Auto-Man-Client-Reg[Automatic and Manual {project_name} Client Registration Methods] for more information on the differences between OpenID-Connect and SAML clients.
====

==== Preparing {project_name} Authentication for OpenShift Deployment
Log in to the OpenShift CLI with a user that holds the _cluster:admin_ role.

. Create a new project:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc new-project eap-app-demo
----
//. Create a service account to be used for the {project_name} deployment:
//+
//[source,bash,subs="attributes+,macros+"]
//----
//$ oc create serviceaccount eap-service-account
//----
. Add the `view` role to the link:https://docs.openshift.com/container-platform/latest/dev_guide/service_accounts.html#default-service-accounts-and-roles[`default`] service account. This enables the service account to view all the resources in the `eap-app-demo` namespace, which is necessary for managing the cluster.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc policy add-role-to-user view system:serviceaccount:$(oc project -q):default
----
. The EAP template requires an xref:Configuring-Keystores[SSL keystore and a JGroups keystore]. +
This example uses `keytool`, a package included with the Java Development Kit, to generate self-signed certificates for these keystores. The following commands will prompt for passwords. +
.. Generate a secure key for the SSL keystore:
+
[source,bash,subs="attributes+,macros+"]
----
$ keytool -genkeypair -alias https -storetype JKS -keystore eapkeystore.jks
----
.. Generate a secure key for the JGroups keystore:
+
[source,bash,subs="attributes+,macros+"]
----
$ keytool -genseckey -alias jgroups -storetype JCEKS -keystore eapjgroups.jceks
----
. Generate the EAP for OpenShift secrets with the SSL and JGroup keystore files:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc secret new eap-ssl-secret eapkeystore.jks
$ oc secret new eap-jgroup-secret eapjgroups.jceks
----
. Add the EAP secret to the `default` service account:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc secrets link default eap-ssl-secret eap-jgroup-secret
----

==== Preparing the {project_name} Credentials
Log in to the encrypted {project_name} web server at *$$https://secure-sso-$$_<project-name>_._<hostname>_/auth/admin* using the xref:sso-administrator-setup[administrator account] created during the {project_name} deployment.

*Create a Realm*

. Hover your cursor over the realm namespace at the top of the sidebar and click*Add Realm*.
. Enter a realm name (this example uses _eap-demo_) and click *Create*.

*Copy the Public Key*

In the newly created _eap-demo_ realm, click the *Keys* tab and copy the generated public key. This example uses the variable _<realm-public-key>_ for brevity. This is used later to deploy the {project_name}-enabled JBoss EAP image.

*Create a Role*

Create a role in {project_name} with a name that corresponds to the JEE role defined in the *web.xml* of the example EAP application. This role is assigned to an {project_name} _application user_ to authenticate access to user applications.

. Click *Roles* in the *Configure* sidebar to list the roles for this realm. This is a new realm, so there should only be the default _offline_access_ role.
. Click *Add Role*.
. Enter the role name (this example uses the role _eap-user-role_) and click *Save*.

*Create Users and Assign Roles*

Create two users:
- Assign the _realm management user_ the *realm-management* roles to handle automatic {project_name} client registration in the {project_name} server.
- Assign the _application user_ the JEE role, created in the previous step, to authenticate access to user applications.

Create the _realm management user_:

. Click *Users* in the *Manage* sidebar to view the user information for the realm.
. Click *Add User*.
. Enter a valid *Username* (this example uses the user _eap-mgmt-user_) and click *Save*.
. Edit the user configuration. Click the *Credentials* tab in the user space and enter a password for the user. After the password has been confirmed you can click *Reset Password* to set the user password. A pop-up window prompts for additional confirmation.
. Click *Role Mappings* to list the realm and client role configuration. In the *Client Roles* drop-down menu, select *realm-management* and add all of the available roles to the user. This provides the user {project_name} server rights that can be used by the JBoss EAP image to create clients.

Create the _application user_:

. Click *Users* in the *Manage* sidebar to view the user information for the realm.
. Click *Add User*.
. Enter a valid *Username* and any additional optional information for the _application user_ and click *Save*.
. Edit the user configuration. Click the *Credentials* tab in the user space and enter a password for the user. After the password has been confirmed you can click *Reset Password* to set the user password. A pop-up window prompts for additional confirmation.
. Click *Role Mappings* to list the realm and client role configuration. In *Available Roles*, add the role created earlier.

==== Deploy the {project_name}-enabled JBoss EAP Image

. Return to the OpenShift web console and click *Add to project* to list the default image streams and templates.
. Use the *Filter by keyword* search bar to limit the list to those that match _sso_. You may need to click *See all* to show the desired application template.
. Select the *_eap71-sso-s2i_* image to list all of the deployment parameters. Include the following {project_name} parameters to configure the {project_name} credentials during the EAP build:
+
[cols="2*", options="header"]
|===
|Variable
|Example Value
|*_APPLICATION_NAME_*
|_sso_

|*_HOSTNAME_HTTPS_*
|_secure-sample-jsp.eap-app-demo.openshift32.example.com_

|*_HOSTNAME_HTTP_*
|_sample-jsp.eap-app-demo.openshift32.example.com_

|*_SOURCE_REPOSITORY_URL_*
|_$$https://repository-example.com/developer/application$$_

|*_SSO_URL_*
|_$$https://secure-sso-sso-app-demo.openshift32.example.com/auth$$_

|*_SSO_REALM_*
|_eap-demo_

|*_SSO_USERNAME_*
|_eap-mgmt-user_

|*_SSO_PASSWORD_*
| _password_

|*_SSO_PUBLIC_KEY_*
|_<realm-public-key>_

|*_HTTPS_KEYSTORE_*
|_eapkeystore.jks_

|*_HTTPS_PASSWORD_*
|_password_

|*_HTTPS_SECRET_*
|_eap-ssl-secret_

|*_JGROUPS_ENCRYPT_KEYSTORE_*
|_eapjgroups.jceks_

|*_JGROUPS_ENCRYPT_PASSWORD_*
|_password_

|*_JGROUPS_ENCRYPT_SECRET_*
|_eap-jgroup-secret_
|===
. Click *Create* to deploy the JBoss EAP image.

It may take several minutes for the JBoss EAP image to deploy.

==== Log in to the JBoss EAP Server Using {project_name}

. Access the JBoss EAP application server and click *Login*. You are redirected to the {project_name} login.
. Log in using the {project_name} user created in the example. You are authenticated against the {project_name} server and returned to the JBoss EAP application server.


[[Example-EAP-Manual]]
=== Example Workflow: Manually Registering EAP Application in {project_name} with SAML Client
This follows on from xref:Example-Deploying-SSO[Example Workflow: Preparing and Deploying the {project_openshift_product_name} image], in which {project_name} was deployed on OpenShift.

This example prepares {project_name} realm, role, and user credentials for an EAP project and configures an EAP for OpenShift deployment. Once deployed, the {project_name} user can be used to authenticate and access JBoss EAP.

[NOTE]
====
This example uses a SAML client but an OpenID-Connect client could also be used. See xref:../advanced_concepts/advanced_concepts.adoc#SSO-Clients[{project_name} Clients] and xref:../advanced_concepts/advanced_concepts.adoc#Auto-Man-Client-Reg[Automatic and Manual {project_name} Client Registration Methods] for more information on the differences between SAML and OpenID-Connect clients.
====

==== Preparing the {project_name} Credentials
Log in to the encrypted {project_name} web server at *$$https://secure-sso-$$_<project-name>_._<hostname>_/auth/admin* using the xref:sso-administrator-setup[administrator account] created during the {project_name} deployment.

*Create a Realm*

. Hover your cursor over the realm namespace (default is *Master*) at the top of the sidebar and click *Add Realm*.
. Enter a realm name (this example uses _saml-demo_) and click *Create*.

*Copy the Public Key*

In the newly created _saml-demo_ realm, click the *Keys* tab and copy the generated public key. This example uses the variable _realm-public-key_ for brevity. This is needed later to deploy the {project_name}-enabled JBoss EAP image.

*Create a Role*

Create a role in {project_name} with a name that corresponds to the JEE role defined in the *web.xml* of the example EAP application. This role will be assigned to an {project_name} _application user_ to authenticate access to user applications.

. Click *Roles* in the *Configure* sidebar to list the roles for this realm. This is a new realm, so there should only be the default _offline_access_ role.
. Click *Add Role*.
. Enter the role name (this example uses the role _saml-user-role_) and click *Save*.

*Create Users and Assign Roles*

Create two users:
- Assign the _realm management user_ the *realm-management* roles to handle automatic {project_name} client registration in the {project_name} server.
- Assign the _application user_ the JEE role, created in the previous step, to authenticate access to user applications.

Create the _realm management user_:

. Click *Users* in the *Manage* sidebar to view the user information for the realm.
. Click *Add User*.
. Enter a valid *Username* (this example uses the user _app-mgmt-user_) and click *Save*.
. Edit the user configuration. Click the *Credentials* tab in the user space and enter a password for the user. After the password has been confirmed you can click *Reset Password* to set the user password. A pop-up window prompts for additional confirmation.
////
Need for the SAML?
. Click *Role Mappings* to list the realm and client role configuration. In the *Client Roles* drop-down menu, select *realm-management* and add all of the available roles to the user. This provides the user {project_name} server rights that can be used by the JBoss EAP image to create clients.
////

Create the _application user_:

. Click *Users* in the *Manage* sidebar to view the user information for the realm.
. Click *Add User*.
. Enter a valid *Username* and any additional optional information for the _application user_ and click *Save*.
. Edit the user configuration. Click the *Credentials* tab in the user space and enter a password for the user. After the password has been confirmed you can click *Reset Password* to set the user password. A pop-up window prompts for additional confirmation.
. Click *Role Mappings* to list the realm and client role configuration. In *Available Roles*, add the role created earlier.

*Create and Configure a SAML Client*:

Clients are {project_name} entities that request user authentication. This example configures a SAML client to handle authentication for the EAP application. This section saves two files, *keystore.jks* and *keycloak-saml-subsystem.xml* that are needed later in the procedure.

Create the SAML Client:

. Click *Clients* in the *Configure* sidebar to list the clients in the realm. Click *Create*.
. Enter a valid *Client ID*. This example uses _sso-saml-demo_.
. In the *Client Protocol* drop-down menu, select *saml*.
. Enter the *Root URL* for the application. This example uses _$$https://demoapp-eap-app-demo.openshift32.example.com$$_.
. Click *Save*.

Configure the SAML Client:

In the *Settings* tab, set the *Root URL* and the *Valid Redirect URLs* for the new *_sso-saml-demo_* client:

. For the *Root URL*, enter the same address used when creating the client. This example uses _$$https://demoapp-eap-app-demo.openshift32.example.com$$_.
. For the *Valid Redirect URLs*, enter an address for users to be redirected to at when they log in or out. This example uses a redirect address relative to the root _$$https://demoapp-eap-app-demo.openshift32.example.com/*$$_.

Export the SAML Keys:

. Click the *SAML Keys* tab in the _sso-saml-demo_ client space and click *Export*.
. For this example, leave the *Archive Format* as *JKS*. This example uses the default *Key Alias* of _sso-saml-demo_ and default *Realm Certificate Alias* of _saml-demo_.
. Enter the *Key Password* and the *Store Password*. This example uses _password_ for both.
. Click *Download* and save the *keystore-saml.jks* file for use later.
. Click the *_sso-saml-demo_* client to return to the client space ready for the next step.

Download the Client Adapter:

. Click *Installation*.
. Use the *Format Option* drop-down menu to select a format. This example uses *Keycloak SAML Wildfly/JBoss Subsystem*.
. Click *Download* and save the file *keycloak-saml-subsystem.xml*.

The *keystore-saml.jks* will be used with the other EAP keystores in the next section to create an OpenShift secret for the EAP application project. Copy the *keystore-saml.jks* file to an OpenShift node. +
The *keycloak-saml-subsystem.xml* will be modified and used in the application deployment. Copy it into the */configuration* folder of the application as *secure-saml-deployments*.

==== Preparing {project_name} Authentication for OpenShift Deployment
Log in to the OpenShift CLI with a user that holds the _cluster:admin_ role.

. Create a new project:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc new-project eap-app-demo
----
//. Create a service account to be used for the SSO deployment:
//+
//[source,bash,subs="attributes+,macros+"]
//----
//$ oc create serviceaccount app-service-account
//----
. Add the `view` role to the link:https://docs.openshift.com/container-platform/latest/dev_guide/service_accounts.html#default-service-accounts-and-roles[`default`] service account. This enables the service account to view all the resources in the `eap-app-demo` namespace, which is necessary for managing the cluster.
+
[source,bash,subs="attributes+,macros+"]
----
$ oc policy add-role-to-user view system:serviceaccount:$(oc project -q):default
----
+
. The EAP template requires an xref:Configuring-Keystores[SSL keystore and a JGroups keystore]. +
This example uses `keytool`, a package included with the Java Development Kit, to generate self-signed certificates for these keystores. The following commands will prompt for passwords. +
.. Generate a secure key for the SSL keystore:
+
[source,bash,subs="attributes+,macros+"]
----
$ keytool -genkeypair -alias https -storetype JKS -keystore eapkeystore.jks
----
.. Generate a secure key for the JGroups keystore:
+
[source,bash,subs="attributes+,macros+"]
----
$ keytool -genseckey -alias jgroups -storetype JCEKS -keystore eapjgroups.jceks
----
. Generate the EAP for OpenShift secrets with the SSL and JGroup keystore files:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc secret new eap-ssl-secret eapkeystore.jks
$ oc secret new eap-jgroup-secret eapjgroups.jceks
----
. Add the EAP application secret to the EAP service account created earlier:
+
[source,bash,subs="attributes+,macros+"]
----
$ oc secrets link default eap-ssl-secret eap-jgroup-secret
----

[[modified-saml-xml]]
==== Modifying the *secure-saml-deployments* File

The *keycloak-saml-subsystem.xml*, exported from the {project_name} client in a previous section, should have been copied into the */configuration* folder of the application and renamed *secure-saml-deployments*. EAP searches for this file when it starts and copies it to the *standalone-openshift.xml* file inside the {project_name} SAML adapter configuration.

. Open the */configuration/secure-saml-deployments* file in a text editor.
. Replace the *YOUR-WAR.war* value of the *secure-deployment name* tag with the application *.war* file. This example uses _sso-saml-demo.war_.
. Replace the *SPECIFY YOUR LOGOUT PAGE!* value of the *logout page* tag with the url to redirect users when they log out of the application. This example uses */index.jsp*.
. Delete the *<PrivateKeyPem>* and *<CertificatePem>* tags and keys and replace it with keystore information:
+
[source,bash,subs="attributes+,macros+"]
----
...
<Keys>
  <Key signing="true">
    <KeyStore file= "/etc/eap-secret-volume/keystore-saml.jks" password="password">
      <PrivateKey alias="sso-saml-demo" password="password"/>
      <Certificate alias="sso-saml-demo"/>
    </KeyStore>
  </Key>
</Keys>
----
+
The mount path of the *keystore-saml.jks* (in this example *_/etc/eap-secret-volume/keystore-saml.jks_*) can be specified in the application template with the parameter *EAP_HTTPS_KEYSTORE_DIR*. +
The aliases and passwords for the *PrivateKey* and the *Certificate* were configured when the SAML Keys were exported from the {project_name} client.
. Delete the second *<CertificatePem>* tag and key and replace it with the the realm certificate information:
+
[source,bash,subs="attributes+,macros+"]
----
...
<Keys>
  <Key signing="true">
    <KeyStore file="/etc/eap-secret-volume/keystore-saml.jks" password="password">
      <Certificate alias="saml-demo"/>
    </KeyStore>
  </Key>
</Keys>
...
----
+
The certificate alias and password were configured when the SAML Keys were exported from the {project_name} client.
. Save and close the */configuration/secure-saml-deployments* file.

==== Configuring SAML Client Registration in the Application *web.xml*

The client type must also be specified by the *<auth-method>* key in the application *web.xml*. This file is read by the image at deployment.

Open the application *web.xml* file and ensure it includes the following:
[source,bash,subs="attributes+,macros+"]
----
...
<login-config>
  <auth-method>KEYCLOAK-SAML</auth-method>
</login-config>
...
----

==== Deploying the Application

You do not need to include any {project_name} configuration for the image because that has been configured in the application itself. Navigating to the application login page redirects you to the {project_name} login. Log in to the application through {project_name} using the _application user_ user created earlier.