Questo contenuto non è disponibile nella lingua selezionata.
Chapter 8. Template-based broker deployment examples
Prerequisites
- These procedures assume an OpenShift Container Platform instance similar to that created in OpenShift Container Platform Getting Started.
- In the AMQ Broker application templates, the values of the AMQ_USER, AMQ_PASSWORD, AMQ_CLUSTER_USER, AMQ_CLUSTER_PASSWORD, AMQ_TRUSTSTORE_PASSWORD, and AMQ_KEYSTORE_PASSWORD environment variables are stored in a secret. To learn more about using and modifying these environment variables when you deploy a template in any of tutorials that follow, see About sensitive credentials.
The following procedures example how to use application templates to create various deployments of brokers.
8.1. Deploying a basic broker with SSL
Deploy a basic broker that is ephemeral and supports SSL.
8.1.1. Deploying the image and template
Prerequisites
- This tutorial builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker tutorial is recommended.
Procedure
- Navigate to the OpenShift web console and log in.
-
Select the
amq-demo
project space. - Click Add to Project > Browse Catalog to list all of the default image streams and templates.
-
Use the Filter search bar to limit the list to those that match
amq
. You might need to click See all to show the desired application template. -
Select the
amq-broker-78-ssl
template which is labeledRed Hat AMQ Broker 7.8 (Ephemeral, with SSL)
. Set the following values in the configuration and click Create.
Table 8.1. Example template Environment variable Display Name Value Description AMQ_PROTOCOL
AMQ Protocols
openwire,amqp,stomp,mqtt,hornetq
The protocols to be accepted by the broker
AMQ_QUEUES
Queues
demoQueue
Creates an anycast queue called demoQueue
AMQ_ADDRESSES
Addresses
demoTopic
Creates an address (or topic) called demoTopic. By default, this address has no assigned routing type.
AMQ_USER
AMQ Username
amq-demo-user
The username the client uses
AMQ_PASSWORD
AMQ Password
password
The password the client uses with the username
AMQ_TRUSTSTORE
Trust Store Filename
broker.ts
The SSL truststore file name
AMQ_TRUSTSTORE_PASSWORD
Truststore Password
password
The password used when creating the Truststore
AMQ_KEYSTORE
AMQ Keystore Filename
broker.ks
The SSL keystore file name
AMQ_KEYSTORE_PASSWORD
AMQ Keystore Password
password
The password used when creating the Keystore
8.1.2. Deploying the application
After creating the application, deploy it to create a Pod and start the broker.
Procedure
- Click Deployments in the OpenShift Container Platform web console.
- Click the broker-amq deployment.
- Click Deploy to deploy the application.
Click the broker Pod and then click the Logs tab to verify the state of the broker.
If the broker logs have not loaded, and the Pod status shows
ErrImagePull
orImagePullBackOff
, your deployment configuration was not able to directly pull the specified broker image from the Red Hat Container Registry. In this case, edit your deployment configuration to reference the correct broker image name and the image pull secret name associated with the account used for authentication in the Red Hat Container Registry. Then, you can import the broker image and start the broker. To do this, complete steps similar to those in Deploying and starting the broker application.
8.1.3. Creating a Route
Create a Route for the broker so that clients outside of OpenShift Container Platform can connect using SSL. By default, the secured broker protocols are available through the 61617/TCP port. In addition, there are SSL and non-SSL ports exposed on the broker Pod for each messaging protocol that the broker supports. However, external client cannot connect directly to these ports on the broker. Instead, external clients connect to OpenShift via the Openshift router, which determines how to forward traffic to the appropriate port on the broker Pod.
If you scale your deployment up to multiple brokers in a cluster, you must manually create a Service and a Route for each broker, and then use each Service-and-Route combination to direct a given client to a given broker, or broker list. For an example of configuring multiple Services and Routes to connect clustered brokers to their own instances of the AMQ Broker management console, see Creating Routes for the AMQ Broker management console.
Prerequisites
- Before creating an SSL Route, you should understand how external clients use this Route to connect to the broker. For more information, see Creating an SSL Route.
Procedure
-
Click
. -
Click
. - To display the TLS parameters, select the Secure route check box.
- From the TLS Termination drop-down menu, choose Passthrough. This selection relays all communication to AMQ Broker without the OpenShift router decrypting and resending it.
To view the Route, click Routes. For example:
https://broker-amq-tcp-amq-demo.router.default.svc.cluster.local
This hostname will be used by external clients to connect to the broker using SSL with SNI.
Additional resources
- For more information about creating SSL Routes, see Creating an SSL Route.
- For more information on Routes in the OpenShift Container Platform, see Routes.
8.2. Deploying a basic broker with persistence and SSL
Deploy a persistent broker that supports SSL. When a broker needs persistence, the broker is deployed as a StatefulSet and stores messaging data on a persistent volume associated with the broker Pod via a persistent volume claim. When a broker Pod is created, it uses storage that remains in the event that you shut down the Pod, or if the Pod shuts down unexpectedly. This configuration means that messages are not lost, as they would be with a standard deployment.
Prerequisites
- This tutorial builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker tutorial is recommended.
- You must have sufficient persistent storage provisioned to your OpenShift cluster to associate with your broker Pod via a persistent volume claim. For more information, see Understanding persistent storage (OpenShift Container Platform 4.5)
8.2.1. Deploy the image and template
Procedure
- Navigate to the OpenShift web console and log in.
-
Select the
amq-demo
project space. -
Click
to list all of the default image streams and templates. -
Use the Filter search bar to limit the list to those that match
amq
. You might need to click See all to show the desired application template. -
Select the
amq-broker-78-persistence-ssl
template, which is labelledRed Hat AMQ Broker 7.8 (Persistence, with SSL)
. Set the following values in the configuration and click create.
Table 8.2. Example template Environment variable Display Name Value Description AMQ_PROTOCOL
AMQ Protocols
openwire,amqp,stomp,mqtt,hornetq
The protocols to be accepted by the broker
AMQ_QUEUES
Queues
demoQueue
Creates an anycast queue called demoQueue
AMQ_ADDRESSES
Addresses
demoTopic
Creates an address (or topic) called demoTopic. By default, this address has no assigned routing type.
VOLUME_CAPACITY
AMQ Volume Size
1Gi
The persistent volume size created for the journal
AMQ_USER
AMQ Username
amq-demo-user
The username the client uses
AMQ_PASSWORD
AMQ Password
password
The password the client uses with the username
AMQ_TRUSTSTORE
Trust Store Filename
broker.ts
The SSL truststore file name
AMQ_TRUSTSTORE_PASSWORD
Truststore Password
password
The password used when creating the Truststore
AMQ_KEYSTORE
AMQ Keystore Filename
broker.ks
The SSL keystore file name
AMQ_KEYSTORE_PASSWORD
AMQ Keystore Password
password
The password used when creating the Keystore
8.2.2. Deploy the application
Once the application has been created it needs to be deployed. Deploying the application creates a Pod and starts the broker.
Procedure
- Click StatefulSets in the OpenShift Container Platform web console.
- Click the broker-amq deployment.
- Click Deploy to deploy the application.
Click the broker Pod and then click the Logs tab to verify the state of the broker. You should see the queue created via the template.
If the broker logs have not loaded, and the Pod status shows
ErrImagePull
orImagePullBackOff
, your configuration was not able to directly pull the specified broker image from the Red Hat Container Registry. In this case, edit your deployment configuration to reference the correct broker image name and the image pull secret name associated with the account used for authentication in the Red Hat Container Registry. Then, you can import the broker image and start the broker. To do this, complete steps similar to those in Deploying and starting the broker application.Click the Terminal tab to access a shell where you can use the CLI to send some messages.
sh-4.2$ ./broker/bin/artemis producer --destination queue://demoQueue Producer ActiveMQQueue[demoQueue], thread=0 Started to calculate elapsed time ... Producer ActiveMQQueue[demoQueue], thread=0 Produced: 1000 messages Producer ActiveMQQueue[demoQueue], thread=0 Elapsed time in second : 4 s Producer ActiveMQQueue[demoQueue], thread=0 Elapsed time in milli second : 4584 milli seconds sh-4.2$ ./broker/bin/artemis consumer --destination queue://demoQueue Consumer:: filter = null Consumer ActiveMQQueue[demoQueue], thread=0 wait until 1000 messages are consumed Received 1000 Consumer ActiveMQQueue[demoQueue], thread=0 Consumed: 1000 messages Consumer ActiveMQQueue[demoQueue], thread=0 Consumer thread finished
Alternatively, use the OpenShift client to access the shell using the Pod name, as shown in the following example.
// Get the Pod names and internal IP Addresses oc get pods -o wide // Access a broker Pod by name oc rsh <broker-pod-name>
Now scale down the broker using the oc command.
$ oc scale statefulset broker-amq --replicas=0 statefulset "broker-amq" scaled
You can use the console to check that the Pod count is 0
Now scale the broker back up to
1
.$ oc scale statefulset broker-amq --replicas=1 statefulset "broker-amq" scaled
Consume the messages again by using the terminal. For example:
sh-4.2$ broker/bin/artemis consumer --destination queue://demoQueue Consumer:: filter = null Consumer ActiveMQQueue[demoQueue], thread=0 wait until 1000 messages are consumed Received 1000 Consumer ActiveMQQueue[demoQueue], thread=0 Consumed: 1000 messages Consumer ActiveMQQueue[demoQueue], thread=0 Consumer thread finished
Additional resources
- For more information on managing stateful applications, see StatefulSets (external).
8.2.3. Creating a Route
Create a Route for the broker so that clients outside of OpenShift Container Platform can connect using SSL. By default, the broker protocols are available through the 61617/TCP port.
If you scale your deployment up to multiple brokers in a cluster, you must manually create a Service and a Route for each broker, and then use each Service-and-Route combination to direct a given client to a given broker, or broker list. For an example of configuring multiple Services and Routes to connect clustered brokers to their own instances of the AMQ Broker management console, see Creating Routes for the AMQ Broker management console.
Prerequisites
- Before creating an SSL Route, you should understand how external clients use this Route to connect to the broker. For more information, see Creating an SSL Route.
Procedure
-
Click
. -
Click
. - To display the TLS parameters, select the Secure route check box.
- From the TLS Termination drop-down menu, choose Passthrough. This selection relays all communication to AMQ Broker without the OpenShift router decrypting and resending it.
To view the Route, click Routes. For example:
https://broker-amq-tcp-amq-demo.router.default.svc.cluster.local
This hostname will be used by external clients to connect to the broker using SSL with SNI.
Additional resources
- For more information on Routes in the OpenShift Container Platform, see Routes.
8.3. Deploying a set of clustered brokers
Deploy a clustered set of brokers where each broker runs in its own Pod.
8.3.1. Distributing messages
Message distribution is configured to use ON_DEMAND. This means that when messages arrive at a clustered broker, the messages are distributed in a round-robin fashion to any broker that has consumers.
This message distribution policy safeguards against messages getting stuck on a specific broker while a consumer, connected either directly or through the OpenShift router, is connected to a different broker.
The redistribution delay is zero by default. If a message is on a queue that has no consumers, it will be redistributed to another broker.
When redistribution is enabled, messages can be delivered out of order.
8.3.2. Deploy the image and template
Prerequisites
- This procedure builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker tutorial is recommended.
Procedure
- Navigate to the OpenShift web console and log in.
-
Select the
amq-demo
project space. - Click Add to Project > Browse catalog to list all of the default image streams and templates
-
Use the Filter search bar to limit the list to those that match
amq
. Click See all to show the desired application template. -
Select the
amq-broker-78-persistence-clustered
template which is labeledRed Hat AMQ Broker 7.8 (no SSL, clustered)
. Set the following values in the configuration and click create.
Table 8.3. Example template Environment variable Display Name Value Description AMQ_PROTOCOL
AMQ Protocols
openwire,amqp,stomp,mqtt,hornetq
The protocols to be accepted by the broker
AMQ_QUEUES
Queues
demoQueue
Creates an anycast queue called demoQueue
AMQ_ADDRESSES
Addresses
demoTopic
Creates an address (or topic) called demoTopic. By default, this address has no assigned routing type.
VOLUME_CAPACITY
AMQ Volume Size
1Gi
The persistent volume size created for the journal
AMQ_CLUSTERED
Clustered
true
This needs to be true to ensure the brokers cluster
AMQ_CLUSTER_USER
cluster user
generated
The username the brokers use to connect with each other
AMQ_CLUSTER_PASSWORD
cluster password
generated
The password the brokers use to connect with each other
AMQ_USER
AMQ Username
amq-demo-user
The username the client uses
AMQ_PASSWORD
AMQ Password
password
The password the client uses with the username
8.3.3. Deploying the application
Once the application has been created it needs to be deployed. Deploying the application creates a Pod and starts the broker.
Procedure
- Click StatefulSets in the OpenShift Container Platform web console.
- Click the broker-amq deployment.
Click Deploy to deploy the application.
NoteThe default number of replicas for a clustered template is 0. You should not see any Pods.
Scale up the Pods to three to create a cluster of brokers.
$ oc scale statefulset broker-amq --replicas=3 statefulset "broker-amq" scaled
Check that there are three Pods running.
$ oc get pods NAME READY STATUS RESTARTS AGE broker-amq-0 1/1 Running 0 33m broker-amq-1 1/1 Running 0 33m broker-amq-2 1/1 Running 0 29m
-
If the Pod status shows
ErrImagePull
orImagePullBackOff
, your deployment was not able to directly pull the specified broker image from the Red Hat Container Registry. In this case, edit your StatefulSet to reference the correct broker image name and the image pull secret name associated with the account used for authentication in the Red Hat Container Registry. Then, you can import the broker image and start the brokers. To do this, complete steps similar to those in Deploying and starting the broker application. Verify that the brokers have clustered with the new Pod by checking the logs.
$ oc logs broker-amq-2
This shows the logs of the new broker and an entry for a clustered bridge created between the brokers:
2018-08-29 07:43:55,779 INFO [org.apache.activemq.artemis.core.server] AMQ221027: Bridge ClusterConnectionBridge@1b0e9e9d [name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, queue=QueueImpl[name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c], temp=false]@5e0c0398 targetConnector=ServerLocatorImpl (identity=(Cluster-connection-bridge::ClusterConnectionBridge@1b0e9e9d [name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, queue=QueueImpl[name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c], temp=false]@5e0c0398 targetConnector=ServerLocatorImpl [initialConnectors=[TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-110], discoveryGroupConfiguration=null]]::ClusterConnectionImpl@806813022[nodeUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c, connector=TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-108, address=, server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c])) [initialConnectors=[TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-110], discoveryGroupConfiguration=null]] is connected
8.3.4. Creating Routes for the AMQ Broker management console
The clustering templates do not expose the AMQ Broker management console by default. This is because the OpenShift proxy performs load balancing across each broker in the cluster and it would not be possible to control which broker console is connected at a given time.
The following example procedure shows how to configure each broker in the cluster to connect to its own management console instance. You do this by creating a dedicated Service-and-Route combination for each broker Pod in the cluster.
Prerequisites
- You have already deployed a clustered set of brokers, where each broker runs in its own Pod. See Deploying a set of clustered brokers.
Procedure
Create a regular Service for each Pod in the cluster, using a StatefulSet selector to select between Pods. To do this, deploy a Service template, in
.yaml
format, that looks like the following:apiVersion: v1 kind: Service metadata: annotations: description: 'Service for the management console of broker pod XXXX' labels: app: application2 application: application2 template: amq-broker-78-persistence-clustered name: amq2-amq-console-XXXX namespace: amq75-p-c-ssl-2 spec: ports: - name: console-jolokia port: 8161 protocol: TCP targetPort: 8161 selector: deploymentConfig: application2-amq statefulset.kubernetes.io/pod-name: application2-amq-XXXX type: ClusterIP
In the preceding template, replace
XXXX
with the ordinal value of the broker Pod you want to associate with the Service. For example, to associate the Service with the first Pod in the cluster, setXXXX
to0
. To associate the Service with the second Pod, setXXXX
to1
, and so on.Save and deploy an instance of the template for each broker Pod in your cluster.
NoteIn the example template shown above, the selector uses the Kubernetes-defined Pod name.
Create a Route for each broker Pod, so that the AMQ Broker management console can connect to the Pod.
Click
. The Edit Route page opens.
-
In the Services drop-down menu, select the previously created broker Service that you want to associate the Route with, for example,
amq2-amq-console-0
. -
Set Target Port to
8161
, to enable access for the AMQ Broker management console. To display the TLS parameters, select the Secure route check box.
From the TLS Termination drop-down menu, choose Passthrough.
This selection relays all communication to AMQ Broker without the OpenShift router decrypting and resending it.
Click Create.
When you create a Route associated with one of broker Pods, the resulting
.yaml
file includes lines that look like the following:spec: host: amq2-amq-console-0-amq75-p-c-2.apps-ocp311.example.com port: targetPort: console-jolokia tls: termination: passthrough to: kind: Service name: amq2-amq-console-0 weight: 100 wildcardPolicy: None
-
In the Services drop-down menu, select the previously created broker Service that you want to associate the Route with, for example,
- To access the management console for a specific broker instance, copy the host URL shown above to a web browser.
Additional resources
- For more information on the clustering of brokers see Configuring message redistribution.
8.4. Deploying a set of clustered SSL brokers
Deploy a clustered set of brokers, where each broker runs in its own Pod and the broker is configured to accept connections using SSL.
8.4.1. Distributing messages
Message distribution is configured to use ON_DEMAND. This means that when messages arrive at a clustered broker, the messages are distributed in a round-robin fashion to any broker that has consumers.
This message distribution policy safeguards against messages getting stuck on a specific broker while a consumer, connected either directly or through the OpenShift router, is connected to a different broker.
The redistribution delay is non-zero by default. If a message is on a queue that has no consumers, it will be redistributed to another broker.
When redistribution is enabled, messages can be delivered out of order.
8.4.2. Deploying the image and template
Prerequisites
- This procedure builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker example is recommended.
Procedure
- Navigate to the OpenShift web console and log in.
-
Select the
amq-demo
project space. - Click Add to Project > Browse catalog to list all of the default image streams and templates.
-
Use the Filter search bar to limit the list to those that match
amq
. Click See all to show the desired application template. -
Select the
amq-broker-78-persistence-clustered-ssl
template which is labeledRed Hat AMQ Broker 7.8 (SSL, clustered)
. Set the following values in the configuration and click create.
Table 8.4. Example template Environment variable Display Name Value Description AMQ_PROTOCOL
AMQ Protocols
openwire,amqp,stomp,mqtt,hornetq
The protocols to be accepted by the broker
AMQ_QUEUES
Queues
demoQueue
Creates an anycast queue called demoQueue
AMQ_ADDRESSES
Addresses
demoTopic
Creates an address (or topic) called demoTopic. By default, this address has no assigned routing type.
VOLUME_CAPACITY
AMQ Volume Size
1Gi
The persistent volume size created for the journal
AMQ_CLUSTERED
Clustered
true
This needs to be true to ensure the brokers cluster
AMQ_CLUSTER_USER
cluster user
generated
The username the brokers use to connect with each other
AMQ_CLUSTER_PASSWORD
cluster password
generated
The password the brokers use to connect with each other
AMQ_USER
AMQ Username
amq-demo-user
The username the client uses
AMQ_PASSWORD
AMQ Password
password
The password the client uses with the username
AMQ_TRUSTSTORE
Trust Store Filename
broker.ts
The SSL truststore file name
AMQ_TRUSTSTORE_PASSWORD
Truststore Password
password
The password used when creating the Truststore
AMQ_KEYSTORE
AMQ Keystore Filename
broker.ks
The SSL keystore file name
AMQ_KEYSTORE_PASSWORD
AMQ Keystore Password
password
The password used when creating the Keystore
8.4.3. Deploying the application
Deploy after creating the application. Deploying the application creates a Pod and starts the broker.
Procedure
- Click StatefulSets in the OpenShift Container Platform web console.
- Click the broker-amq deployment.
Click Deploy to deploy the application.
NoteThe default number of replicas for a clustered template is
0
, so you will not see any Pods.Scale up the Pods to three to create a cluster of brokers.
$ oc scale statefulset broker-amq --replicas=3 statefulset "broker-amq" scaled
Check that there are three Pods running.
$ oc get pods NAME READY STATUS RESTARTS AGE broker-amq-0 1/1 Running 0 33m broker-amq-1 1/1 Running 0 33m broker-amq-2 1/1 Running 0 29m
-
If the Pod status shows
ErrImagePull
orImagePullBackOff
, your deployment was not able to directly pull the specified broker image from the Red Hat Container Registry. In this case, edit your StatefulSet to reference the correct broker image name and the image pull secret name associated with the account used for authentication in the Red Hat Container Registry. Then, you can import the broker image and start the brokers. To do this, complete steps similar to those in Deploy and start the broker application. Verify the brokers have clustered with the new Pod by checking the logs.
$ oc logs broker-amq-2
This shows all the logs of the new broker and an entry for a clustered bridge created between the brokers, for example:
2018-08-29 07:43:55,779 INFO [org.apache.activemq.artemis.core.server] AMQ221027: Bridge ClusterConnectionBridge@1b0e9e9d [name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, queue=QueueImpl[name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c], temp=false]@5e0c0398 targetConnector=ServerLocatorImpl (identity=(Cluster-connection-bridge::ClusterConnectionBridge@1b0e9e9d [name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, queue=QueueImpl[name=$.artemis.internal.sf.my-cluster.4333c830-ab5f-11e8-afb8-0a580a82006e, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c], temp=false]@5e0c0398 targetConnector=ServerLocatorImpl [initialConnectors=[TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-110], discoveryGroupConfiguration=null]]::ClusterConnectionImpl@806813022[nodeUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c, connector=TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-108, address=, server=ActiveMQServerImpl::serverUUID=9cedb69d-ab5e-11e8-87a4-0a580a82006c])) [initialConnectors=[TransportConfiguration(name=artemis, factory=org-apache-activemq-artemis-core-remoting-impl-netty-NettyConnectorFactory) ?port=61616&host=10-130-0-110], discoveryGroupConfiguration=null]] is connected
Additional resources
- To learn how to configure each broker in the cluster to connect to its own management console instance, see Creating Routes for the AMQ Broker management console.
- For more information about messaging in a broker cluster, see Enabling Message Redistribution.
8.5. Deploying a broker with custom configuration
Deploy a broker with custom configuration. Although functionality can be obtained by using templates, broker configuration can be customized if needed.
Prerequisites
- This tutorial builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker tutorial is recommended.
8.5.1. Deploy the image and template
Procedure
- Navigate to the OpenShift web console and log in.
-
Select the
amq-demo
project space. - Click Add to Project > Browse catalog to list all of the default image streams and templates.
-
Use the Filter search bar to limit results to those that match
amq
. Click See all to show the desired application template. -
Select the
amq-broker-78-custom
template which is labeledRed Hat AMQ Broker 7.8(Ephemeral, no SSL)
. In the configuration, update
broker.xml
with the custom configuration you would like to use. Click Create.NoteUse a text editor to create the broker’s XML configuration. Then, cut and paste confguration details into the
broker.xml
field.NoteOpenShift Container Platform does not use a
ConfigMap
object to store the custom configuration that you specify in thebroker.xml
field, as is common for many applications deployed on this platform. Instead, OpenShift temporarily stores the specified configuration in an environment variable, before transferring the configuration to a standalone file when the broker container starts.
8.5.2. Deploy the application
Once the application has been created it needs to be deployed. Deploying the application creates a Pod and starts the broker.
Procedure
- Click Deployments in the OpenShift Container Platform web console.
- Click the broker-amq deployment
- Click Deploy to deploy the application.
8.6. Basic SSL client example
Implement a client that sends and receives messages from a broker configured to use SSL, using the Qpid JMS client.
Prerequisites
- This tutorial builds upon Preparing a template-based broker deployment.
- Completion of the Deploying a basic broker with SSL tutorial is recommended.
- AMQ JMS Examples
8.6.1. Configuring the client
Create a sample client that can be updated to connect to the SSL broker. The following procedure builds upon AMQ JMS Examples.
Procedure
Add an entry into your /etc/hosts file to map the route name onto the IP address of the OpenShift cluster:
10.0.0.1 broker-amq-tcp-amq-demo.router.default.svc.cluster.local
Update the jndi.properties configuration file to use the route, truststore and keystore created previously, for example:
connectionfactory.myFactoryLookup = amqps://broker-amq-tcp-amq-demo.router.default.svc.cluster.local:8443?transport.keyStoreLocation=<keystore-path>client.ks&transport.keyStorePassword=password&transport.trustStoreLocation=<truststore-path>/client.ts&transport.trustStorePassword=password&transport.verifyHost=false
Update the jndi.properties configuration file to use the queue created earlier.
queue.myDestinationLookup = demoQueue
- Execute the sender client to send a text message.
Execute the receiver client to receive the text message. You should see:
Received message: Message Text!
8.7. External clients using sub-domains example
Expose a clustered set of brokers through a node port and connect to it using the core JMS client. This enables clients to connect to a set of brokers which are configured using the amq-broker-78-persistence-clustered-ssl
template.
8.7.1. Exposing the brokers
Configure the brokers so that the cluster of brokers are externally available and can be connected to directly, bypassing the OpenShift router. This is done by creating a route that exposes each pod using its own hostname.
Prerequisites
Procedure
- Choose import YAML/JSON from Add to Project drop down
Enter the following and click create.
apiVersion: v1 kind: Route metadata: labels: app: broker-amq application: broker-amq name: tcp-ssl spec: port: targetPort: ow-multi-ssl tls: termination: passthrough to: kind: Service name: broker-amq-headless weight: 100 wildcardPolicy: Subdomain host: star.broker-ssl-amq-headless.amq-demo.svc
NoteThe important configuration here is the wildcard policy of
Subdomain
. This allows each broker to be accessible through its own hostname.
8.7.2. Connecting the clients
Create a sample client that can be updated to connect to the SSL broker. The steps in this procedure build upon the AMQ JMS Examples.
Procedure
Add entries into the /etc/hosts file to map the route name onto the actual IP addresses of the brokers:
10.0.0.1 broker-amq-0.broker-ssl-amq-headless.amq-demo.svc broker-amq-1.broker-ssl-amq-headless.amq-demo.svc broker-amq-2.broker-ssl-amq-headless.amq-demo.svc
Update the jndi.properties configuration file to use the route, truststore, and keystore created previously, for example:
connectionfactory.myFactoryLookup = amqps://broker-amq-0.broker-ssl-amq-headless.amq-demo.svc:443?transport.keyStoreLocation=/home/ataylor/projects/jboss-amq-7-broker-openshift-image/client.ks&transport.keyStorePassword=password&transport.trustStoreLocation=/home/ataylor/projects/jboss-amq-7-broker-openshift-image/client.ts&transport.trustStorePassword=password&transport.verifyHost=false
Update the jndi.properties configuration file to use the queue created earlier.
queue.myDestinationLookup = demoQueue
- Execute the sender client code to send a text message.
Execute the receiver client code to receive the text message. You should see:
Received message: Message Text!
Additional resources
- For more information on using the AMQ JMS client, see AMQ JMS Examples.
8.8. External clients using port binding example
Expose a clustered set of brokers through a NodePort and connect to it using the core JMS client. This enables clients that do not support SNI or SSL. It is used with clusters configured using the amq-broker-78-persistence-clustered
template.
8.8.1. Exposing the brokers
Configure the brokers so that the cluster of brokers are externally available and can be connected to directly, bypassing the OpenShift router. This is done by creating a service that uses a NodePort to load balance around the clusters.
Prerequisites
Procedure
- Choose import YAML/JSON from Add to Project drop down.
Enter the following and click create.
apiVersion: v1 kind: Service metadata: annotations: description: The broker's OpenWire port. service.alpha.openshift.io/dependencies: >- [{"name": "broker-amq-amqp", "kind": "Service"},{"name": "broker-amq-mqtt", "kind": "Service"},{"name": "broker-amq-stomp", "kind": "Service"}] creationTimestamp: '2018-08-29T14:46:33Z' labels: application: broker template: amq-broker-78-statefulset-clustered name: broker-external-tcp namespace: amq-demo resourceVersion: '2450312' selfLink: /api/v1/namespaces/amq-demo/services/broker-amq-tcp uid: 52631fa0-ab9a-11e8-9380-c280f77be0d0 spec: externalTrafficPolicy: Cluster ports: - nodePort: 30001 port: 61616 protocol: TCP targetPort: 61616 selector: deploymentConfig: broker-amq sessionAffinity: None type: NodePort status: loadBalancer: {}
NoteThe NodePort configuration is important. The NodePort is the port in which the client will access the brokers and the type is NodePort.
8.8.2. Connecting the clients
Create consumers that are round-robinned around the brokers in the cluster using the AMQ broker CLI.
Procedure
In a terminal create a consumer and attach it to the IP address where OpenShift is running.
artemis consumer --url tcp://<IP_ADDRESS>:30001 --message-count 100 --destination queue://demoQueue
Repeat step 1 twice to start another two consumers.
NoteYou should now have three consumers load balanced across the three brokers.
Create a producer to send messages.
artemis producer --url tcp://<IP_ADDRESS>:30001 --message-count 300 --destination queue://demoQueue
Verify each consumer receives messages.
Consumer:: filter = null Consumer ActiveMQQueue[demoQueue], thread=0 wait until 100 messages are consumed Consumer ActiveMQQueue[demoQueue], thread=0 Consumed: 100 messages Consumer ActiveMQQueue[demoQueue], thread=0 Consumer thread finished