Release Notes for Red Hat AMQ Broker 7.5
Release Notes for AMQ Broker
Chapter 1. Features
This section outlines the major features added in AMQ 7.1 and later.
- Brokers can be fully redundant, allowing a backup broker to take over connections from a primary broker in the event of a failure. Optionally, a shared volume can be configured between the primary and backup broker.
- Multi-site fault tolerance
- You can use specific broker topologies and the Red Hat Ceph Storage software-defined storage platform to ensure continuity of your messaging system during a data center outage. This type of solution is called a multi-site, fault-tolerant architecture. For more information, see Configuring a multi-site, fault-tolerant messaging system.
- Delivery guarantee
- Some messages that a developer sends from his application require a guarantee of delivery guaranteeing that messages get from the application sending them to the application consuming and processing them. If a network link goes down or the broker experiences a power outage while the message is in transit the message broker must resend undelivered messages once it returns online or the network connection is restored.
- Zero persistence
- Brokers can be configured to not persist any messages for scenarios in which no guarantee is required, but high performance is.
- Basic authentication
- Brokers can authenticate clients using basic username and password credentials.
- Role-based authentication
- Brokers can authenticate and authorize clients to specific queues or message addresses based on roles assigned to the client.
- SSL-encrypted connections
- Connections between clients and broker or a broker and another broker are secure and utilize SSL to encrypt the connections.
- Certificate-based authentication
- Brokers can authenticate a client connection or a connection from another broker using certificates.
- Dead connection detection
- Brokers can detect dead connections when a client has exceeded specific lifespan that can be set either on the client’s connection factory or globally on the server.
- Slow consumer detection
- Brokers can detect slow consumers and disconnect them from the server when this feature is enabled.
Refer to the ActiveMQ Artemis project for additional information: http://activemq.apache.org/artemis/index.html
Chapter 2. Enhancements
ENTMQBR-428 - Implement SASL external
AMQ Broker can now validate AMQP connections by using the identity of the connection’s SSL certificate rather than using traditional credentials such as a username and password. This benefits deployments where secure access is managed through SSL certificates.
ENTMQBR-831 - Broker should respond to changes in
AMQ Broker now retains the changes that you make to the
logging.propertiesconfiguration file without restarting the broker.
ENTMQBR-893 - Masking of password does not work with
You can now mask passwords in the JAAS configuration file
login.config. Previously, password masking was not supported for
login.config, which meant that the passwords in the file were stored in plain text.
ENTMQBR-908 - Support masked passwords in
You can now mask passwords in the
broker.xmlconfiguration file. Previously, password masking was not supported for
broker.xml, which meant that the passwords in the file were stored in plain text.
ENTMQBR-935 - Should have Ability to split
broker.xmlinto multiple files and import into
AMQ Broker now supports XML inclusions in the
broker.xmlconfiguration file, which enables you to break
broker.xmlinto separate files. For example, if you have multiple brokers that share the same set of addresses and address settings, you can define the addresses in a separate file, and then include that file in each broker’s
ENTMQBR-945 - Non-persistent messages lost in non-failure scenarios when authorization fails, because delivery mode defaults to asynchronous
AMQ Broker now enables you to configure notifications for loss of non-persistent messages by clients. Obvious failure conditions such as broker stoppage or disconnection of a storage provider can cause this type of loss. However, the default delivery mode for non-persistent messages, which is asynchronous, can also cause message loss in situations such as failed authorization or use of a non-configured queue. You can now use the CompletionListener API to configure notifications for such message-loss events.
ENTMQBR-1002 - Enable the pause method for an address
AMQ Broker 7.5 enables you to pause an address, which pauses all of the queues bound to that address. New queues that you add to this address are also paused until you resume the address. To pause or resume an address, use the Core JMS API to implement the
resume()methods on an
AddressControlinterface. For more information, see AddressControl in the Core JMS API documentation.
ENTMQBR-1056 - Add Support for DB2 v11.latest
The 7.2 version of AMQ Broker adds support for the database DB2, version 11.
ENTMQBR-1063 - Improve DB2 compatibility
AMQ Broker now supports IBM DB2 versions 10.5 and 11.1 for message storage. The following enhancements have been made:
AUTO_INCREMENTis now supported on DB2.
- You can now configure the maximum allowed time difference between the system time and DB2 time. This makes high availability more reliable, because the broker will be shut down automatically if the time difference exceeds the configured value.
- Database tables can now be created externally. Concurrent table creation checking has been improved.
ENTMQBR-1271 - Allow custom etc and data directories to be used
AMQ Broker 7.2 provides greater flexibility in storing a broker instance’s configuration files and data. Previously, a broker instance’s configuration files and data could only be stored in the
data/directories within the broker instance’s directory. Now, you can store these files in any custom directory, including locations outside of the broker instance’s directory.
ENTMQBR-1282 - Support for exclusive consumers
AMQ Broker now supports exclusive queues. An exclusive queue routes messages to a single consumer even if multiple consumers are present. In addition, the consumer that receives the messages will receive them with guaranteed message ordering so that they can be consumed in the same order in which they were produced.
ENTMQBR-1308 - Create an
ActiveMQServerPluginfor logging various broker events
Starting in AMQ Broker 7.2, logging has been improved through a new
LoggingActiveMQServerPlugin. By using this plugin, you can configure highly-informative logging about many different types of events that would otherwise only be available through debug logs.
ENTMQBR-1367 - Support SSL Certificate Revocation List
Starting in AMQ Broker 7.2, you can now specify an SSL certificate revocation list on an acceptor’s URL. If you are using SSL for client communication, this means that you can specify a list of certificates that have been revoked by the issuing certificate authority (CA) before their scheduled expiration date and should no longer be trusted.
ENTMQBR-1370 - Add
userattribute to queue in XML/JMX
When a queue is auto-created it is assigned the user name of the connecting client. This user name is now exposed as metadata on the queue and also in JMX and in the AMQ Console. You can also now configure a user on a pre-defined queue by modifying the
broker.xmlfile. This enhancement makes it possible to understand and track with whom and how queues originated.
ENTMQBR-1385 - Add support for virtual topic consumers
AMQ Broker now supports migrating existing OpenWire clients that use virtual topic subscriptions (an ActiveMQ 5.x-only feature) without requiring any changes to the client. This functionality maps the virtual topic concept to an Artemis fully qualified queue name (FQQN). By using a FQQN, you can create and access a standard Artemis topic subscriber queue.
ENTMQBR-1403 - Support referrals in LDAP login module
AMQ Broker now supports LDAP referrals. If you use LDAP referrals in your LDAP infrastructure, you can now configure the broker to use these referrals to authenticate and authorize users. You can use the broker’s JAAS LDAP login module to control how the broker handles LDAP referrals.
ENTMQBR-1648 - Don’t hardcode openshift.DNS_PING service name
AMQ Broker 7.5 enables you to specify a unique value for the OpenShift DNS ping service for each application template deployment that you make in a given OpenShift project. This enhancement enables you to deploy multiple broker clusters in the same OpenShift project. For more information, see the description of the
PING_SVC_NAMEparameter in Application template parameters.
ENTMQBR-1783 - Add a user and role manager (create/delete/update/list) JMX endpoint in AMQ 7
AMQ Broker 7.3 supports adding, removing, and updating user information by using the management API when you are using the
PropertiesLoginModuleclass. Previously, you could perform these actions only by using the command-line interface.
ENTMQBR-1918 - [AMQ7, critical-analyzer] add extra detail to output for analyzing cause of timeout
The logs for the Critical Analyzer now provide extra information about a slow-performing broker when you enable tracing on the package named
ENTMQBR-1932 - Support AMQ 7 on Java 11
AMQ Broker 7.4 adds runtime support for JDK 11 Java Virtual Machines (JVMs).
ENTMQBR-1974 - Support masked passwords in
AMQ Broker now supports masking the values of the
trust-store-passwordsettings in the
management.xmlfile. These values are used when you secure remote JMX access with SSL.
ENTMQBR-2020 - Track routed and unrouted messages to an address
Address metrics now include information about how many messages the address has received. This number is divided between routed and unrouted messages. Prior to this enhancement, it was impossible to determine if an address had received a message which was not routed, for example, in the case of a JMS topic with no subscriptions.
ENTMQBR-2025 - Allow users to define custom last value queue keys
AMQ Broker 7.4 supports last value queues with custom last value key names. Previously, last value queues used only the default last value key name of
ENTMQBR-2139 - Support MSSQL for database message persistence
AMQ Broker 7.4 supports Microsoft SQL Server 2016 for JDBC-based persistence.
ENTMQBR-2173 - Support a Prometheus plugin or agent
AMQ Broker 7.4 includes a pre-configured plugin for Prometheus. Prometheus is software built to monitor large, scalable systems and store historical runtime data over an extended time period. The AMQ Broker Prometheus plugin exports broker runtime metrics to Prometheus format, enabling you to use Prometheus itself to visualize and run queries on the data.
ENTMQBR-2299 - An option to reject messages without a JMSXUserID
Starting in AMQ Broker 7.4, you can track and log the origins of messages using the
_AMQ_VALIDATED_USERmessage key. You can also configure the broker to reject a message that does not have a user name already populated by the client when it sends the message.
ENTMQBR-2345 - Enhancements to the databaseStoreType
In AMQ Broker 7.4, you can add an encrypted user name and password when configuring a
database-storeelement in your
ENTMQBR-2687 - Add new ActiveMQArtemis CRD version to reflect actual configuration schema
In AMQ Broker 7.5, the Custom Resource Definition (CRD) files used by the AMQ Broker Operator have been redesigned. This enhancement makes the CRDs more consistent with the
broker.xmlconfiguration file that you use to configure on-premise broker deployments. For more information, see Overview of Custom Resource Definitions.
ENTMQBR-2846 - Operator-based deployments larger than one broker are clustered by default
In the 7.5 version of the AMQ Broker Operator, a deployment of two or more brokers is clustered by default. In the 7.4 version of the Operator, you needed to explicitly include a
clusterConfigconfiguration element in the Custom Resource (CR) that you created for your deployment.
ENTMQBR-2921 - Scaledown controller no longer explicitly deployed via Custom Resource
In the 7.5 version of the AMQ Broker Operator, if you intend to enable message migration in your broker deployment, you no longer need to deploy a dedicated Custom Resource (CR) instance for the scaledown controller. Now, if you set
truein your main broker CR, the Operator automatically instantiates and starts the scaledown controller to prepare for message migration. If you set
false, the Operator does not instantiate the scaledown controller. If you do not explicitly set a value for
messageMigration, the Operator sets a default value of
true. In this case, the Operator also automatically instantiates and starts the scaledown controller.
ENTMQBR-3126 - Added AMQ Broker support for running within OpenShift Container Platform on IBM Z
AMQ Broker 7.5 adds broker and Operator container images that support RHEL 8 on IBM Z. This added support enables you to create broker deployments within OpenShift Container Platform on IBM Z. For more information, see Deploying AMQ Broker on OpenShift Container Platform Using an Operator.
Chapter 3. Deprecated features
Starting in 7.3, AMQ Broker no longer ships with the Hawtio dispatch console plugin,
dispatch-hawtio-console.war. Previously, the dispatch console was used to manage AMQ Interconnect. However, AMQ Interconnect now uses its own, standalone web console.
- Starting in 7.5, network pinging is a deprecated feature. Network pinging cannot protect a broker cluster from network isolation issues that can lead to irrecoverable message loss. This feature will be removed in a future release. Red Hat continues to support existing AMQ Broker deployments that use network pinging. However, Red Hat no longer recommends use of network pinging in new deployments. For guidance on configuring a broker cluster for high availability and to avoid network isolation issues, see Implementing high availability.
Chapter 4. Technology preview
Technology Preview features are not supported with Red Hat production service-level agreements (SLAs) and might not be functionally complete. Red Hat does not recommend using them for production. For more information, see Red Hat Technology Preview Features Support Scope.
- Support for federated queues and addresses is available as a Technology Preview feature only. Federation enables you to transmit messages between brokers without the brokers being part of a cluster. For more information, see Federation in the Apache ActiveMQ Artemis documentation.
Chapter 5. Fixed issues
ENTMQBR-742 - JMS Queues are not being auto-deleted for Openwire and AMQP clients
Previously, addresses and queues with
truewere not deleted when AMQ OpenWire JMS clients disconnected from them.
ENTMQBR-781 - [Error Codes] same Error Code(s) defined in different places with different reasons
In previous releases of AMQ Broker, logging error codes were not unique for all broker log messages. The duplicate and conflicting logging error codes are now corrected.
ENTMQBR-882 - Standby slave does not announce replication to master when primary slave is down
This issue occurred in past releases when multiple backup brokers, also referred to as slaves, served a single live (master) broker. In this scenario, if a primary backup broker failed, the secondary backup tried to replicate. But that operation failed, the secondary backup could not take over for the primary backup, and as a result, high-availability was lost. This issue is now resolved.
ENTMQBR-914 - [AMQ7, broker startup ] AMQ224000: Failure in initialization:
java.lang.IllegalStateException: Cannot find queue with id XXXX
In previous versions of AMQ Broker, the broker would fail to start if you had an XA transaction in the prepared state, then deleted the queue that this transaction used, and then attempted to restart the broker.
LDAPLoginModulecannot process referrals
Previously, the LDAP JAAS login module was unable to handle LDAP referrals, which caused authentication and authorization failures.
ENTMQBR-930 - Unable to login with multiple LDAP modules configured
The commit operation of the LDAP JAAS login module would always return
null, resulting in unexpected behavior when multiple instances of the module were configured in the same domain.
ENTMQBR-943 - [AMQ7, Openwire, Compression] consuming Openwire compressed
java.util.zip.DataFormatException: incorrect header check
Previously, when using the OpenWire protocol to send small, compressed ByteMessages that have JMS properties set for a queue, an exception was thrown on the consumer side when it attempted to decompress the message. See the Knowledge Base article on the Red Hat Customer Portal for more details: https://access.redhat.com/solutions/3269061.
ENTMQBR-956 - [AMQ7, AMQP, JMX, DeliveringCount] JMX operation
listDeliveringMessages()returns empty string when
In previous releases, the JMX operation
listDeliveringMessages()did not return messages that were consumed by an AMQ JMS client in a transaction. The transactional messages are now returned correctly when the JMX operation
ENTMQBR-965 - [AMQ7, openwire, exclusive consumer] AMQ6 OpenWire clients using exclusive consumer do not behave the same with an AMQ 7 broker
In AMQ 7, setting the maximum number of consumers for a queue to one did not work the same as for an exclusive consumer in AMQ 6. In AMQ 6, using an exclusive consumer means that one consumer will receive all of the messages. Any other consumers are still registered, just not consuming messages. When the current exclusive consumer leaves the queue, another consumer becomes the exclusive consumer.
Previously in AMQ 7, if you set
max-consumers = 1for a queue, when this limit is reached, the broker returned a
AMQ119200error and did not allow any new consumers to join the queue. This error no longer occurs.
ENTMQBR-966 - Unsettled AMQP messages are lost when Receiver Link is opened on remote cluster member
An issue causing message loss has been fixed in this release. Previously, if messages were sent to a broker using the AMQP address, and the address was not set on the messages, then some of the messages could be lost if they were redistributed.
ENTMQBR-967 - [AMQ 7.1.0 CR1.1] Limit non-ssl connection, handshake-timeout not configurable
Previously, the broker did not disconnect unauthenticated clients. With AMQ Broker 7.2, you can use the configuration parameter
handshake-timeoutto limit the amount of time that an unauthenticated client can remain connected.
ENTMQBR-973 - Incorrect message priority displayed in hawtio console
When viewing messages in AMQ Console, the message priority is now correct. Before, the message priority incorrectly defaulted to
ENTMQBR-1016 - [AMQ7,Hawtio]AMQ 7 hawtio console store users password in browser’s local cache after user get logout
A security issue has been fixed for AMQ Console. Before, if you logged into AMQ Console, the value of the
Passwordfield was visible from local storage in Google Chrome Developer tools.
ENTMQBR-1018 - When live-slave fails-back to master, it turns off everything down, even its console
In high-availability configurations, AMQ Console is now accessible when a slave broker returns control to the master broker. Previously, AMQ Console would become unavailable for the slave broker when it gave control back to the master broker.
ENTMQBR-1030 - Restrict directory listings of Hawtio within the web server configuration
AMQ Console no longer permits access to restricted directory listings.
Client connection failedmessage in broker log with OpenWire client
Previously, when an OpenWire client (ActiveMQ) disconnected cleanly, the broker showed a warning message similar to the following:
Client connection failed, clearing up resources for session ID:localhost-45795-1527065581471-5:1:1
This issue is now resolved.
ENTMQBR-1061 - Exception when using LegacyLDAPSecuritySettingsPlugin
Previously, when using LegacyLDAPSecuritySettingsPlugin, if you specified authorization settings with non-standard, but still valid syntax, you saw an exception. This is because the broker interpreted LDAP search results based on hard-coded key names, not on order. This issue is corrected.
ENTMQBR-1130 - Destinations undeployed when master recovers from outage
When adding destinations to a broker’s configuration file (
broker.xml) at runtime, the destinations are now preserved in the configuration file and reloaded if the broker is restarted. Previously, if you added destinations to a broker’s configuration file, the destinations would not be reloaded when the broker was restarted.
ENTMQBR-1169 - [AMQ 7, Hawtio, large message] Sending very large messages through Hawtio results in an exception
Previously, if you used the Hawtio console to send a message exceeding the broker’s internal buffer size, you saw an exception. This issue is corrected. The broker now automatically converts this type of message to a large message and sends it successfully.
ENTMQBR-1184 - LargeMessage Produced by AMQP Protocol Can Not Be Consumed By AMQP Protocol
In previous releases, if the size of an AMQP JMS Object Message was greater than the value specified for the maximum journal record size, an exception was thrown on the broker and the consumer was not able to receive the message. This issue was caused by a problem in the AMQP large message to core message conversion process.
This issue is fixed and AMQP large messages can be sent and received as usual.
ENTMQBR-1461 - AMQP:
ObjectMessagethat was handled as a Large Message
Previously, if you were using the AMQP protocol with the Qpid JMS client, and you sent a JMS
ObjectMessagethat was also a large message (larger than the
min-large-message-size), an error would occur when the message was consumed. This error no longer occurs.
ENTMQBR-1466 - [3 HA pairs] Slave does not become live after master is killed and isolates itself
The quorum voting protocol has been corrected. Previously, in high-availability configurations consisting of three high-availability pairs with replication, this issue occasionally prevented slave brokers from taking over during a failover event. Instead, the slave broker would become isolated from the broker cluster.
ENTMQBR-1500 - Jolokia read request does not fetch all attributes
Using Jolokia, it is now possible to request all of the attributes for the broker MBean.
ENTMQBR-1548 - Implementation of AMQP interceptor is passing a null RemotingConnection reference
Previously, if you were using the interceptor API with the AMQP protocol, and you implemented the following method, the connection parameter was always null:
public boolean intercept(AMQPMessage message, RemotingConnection connection)
Now, the connection parameter is properly set.
ENTMQBR-1699 - Cannot set the message content when using
Earlier releases of AMQ Broker 7.x did not contain an option to specify the message content when using the
./artemisproducer, as the A-MQ 6.x product did. Now, the
--messageparameter is available and you can use it to specify the content of the message.
ENTMQBR-1757 - AMQ Broker throws ERROR if we update the address ANYCAST to MULTICAST or vice-versa in broker.xml
Previously, if the broker was stopped and the routing type for an existing address with queues was changed in the
broker.xmlfile, the broker would fail to restart. The relevant code for updating the configuration has been modified so that such a configuration change is possible, and even if there is an error deploying the address or queue, the broker will log the error and still start.
ENTMQBR-1816 - Queue#MessageCount negative and messages stop flowing to AMQP consumer
Previously, an AMQP client could send a disposition for a delivered message indicating that the client received the message but didn’t consume it. This behavior could cause the message count on the broker to have a negative value. This issue is now resolved.
ENTMQBR-1878 - [AMQ7, broker shutdown, JMX connector] JVM does not exit if
connectorconfigured in management-context
In previous releases, sometimes the broker did not properly close the management context when stopping the broker. The broker would appear to be stopped, but the JVM process never fully exited. Now there are checks to ensure that the management context is always closed when stopping the broker and the JVM process exits, even when a management context
ENTMQBR-1880 - Hawtio Artemis tab disappears if we use IE
ENTMQBR-1958 - Queue preference changes to display columns not persistent through page refresh
Refreshing the AMQ Console caused changes to the default column lists of tabs to be lost. To maintain your settings, you had to refrain from refreshing the browser window. Now column list preferences are stored and reused each time you access a page and refreshing the browser does not cause settings to be lost.
ENTMQBR-1995 - Enabling TRACE-level logging in the broker leads to
ConcurrentModificationExceptionand failed operations
In previous releases, multiple threads could concurrently modify a message’s properties. This resulted in
WARNmessages in the log when TRACE-level logging was enabled. Now, concurrent access to message properties is done safely and the
WARNmessages in the log related to this issue no longer occur.
ENTMQBR-2011 - Consumer of store-forward internal queues get dropped when syncing large messages
Using temporary destinations in a clustered environment caused messages to get dropped from store and forward queues when clustering messages between brokers. As a result, the broker was not able to distribute messages across cluster nodes and needed to be restarted. Now you can safely use temporary destinations in a clustered environment.
ENTMQBR-2014 - AMQ7:
io.netty.handler.codec.DecoderException: java.lang.NullPointerExceptionin broker log
When a client connected by using an unrecognized protocol, a
NullPointerExceptionwith full stack-trace was logged. Now the AMQ Broker detects and handles unrecognized protocols more gracefully and a proper
ERRORmessage is logged.
ENTMQBR-2024 - Broker ignores client requests for advisory consumers when advisory messages are disabled
In prior releases, when advisory messages were disabled on the broker, the broker ignored client requests to create advisory consumers and made no record of these consumers. If clients attempted to query these consumers, you saw an exception. This issue is corrected. The broker now creates records of advisory consumers when clients request them, and subsequent operations are successful.
ENTMQBR-2052 - Message load balancing off, unable to send messages to queue on other nodes instead of subscribed consumer
In releases prior to AMQ Broker 7.2.3, a recursive call would result in a
StackOverflowErrorerror message in certain conditions when message load balancing was set to
OFF.As a result, the broker would not receive messages sent from clients. In AMQ Broker 7.2.3, a verification step has been added to ensure that a recursive call will not become infinite. Now, brokers receive any message sent by a client even when message load balancing is set to
ENTMQBR-2055 - [Hawtio] console display blank ClientID for the core client
When using releases prior to AMQ Broker 7.2.3 with AMQ Console, the field named Client ID was not populated when viewing a AMQ Core Protocol JMS Client connection for a producer or consumer. Now, the value of Client ID is visible in AMQ Console for AMQ Core Protocol JMS Client connections.
ENTMQBR-2063 - Broker does not send sasl error when user attempts anonymous sasl and GuestLoginModule is not configured
In prior releases, if a client attempted an anonymous login to a broker that did not have guest access enabled, the broker produced an exception. However, this exception was not recognizable to the client as an authentication error. This issue is now resolved.
ENTMQBR-2074 - Retry from DLQ and Expiry Queue is broken on web console (Hawtio)
Previously, the broker ignored messages retried for delivery from the DLQ. These messages instead stayed on the DLQ. This issue is corrected. The broker now correctly adds redelivered messages from the DLQ to the original queue.
ENTMQBR-2085 - [AMQ7, JMSXGroupSeq] Support JMSXGroupSeq -1 to close/reset Groups
Previously, the broker did not remove message groups as expected when a client set JMSXGroupSeq to -1, per ActiveMQ 5. As a result, you could not change a message group once you pinned it to a consumer and broker. This issue is resolved.
ENTMQBR-2086 - [AMQ 7, JMS Selectors] JMS Selectors broken in some cases
With each supported protocol, fields such as
AMQGroupIDwere not properly parsed by the filter parser. As a result, you could not create a filter using any query that used those field names. To correct this issue, wrappers to parse these pseudo-fields were implemented and now filters correctly parse the input.
ENTMQBR-2100 - AMQ 7.2.1: under heavy load: org.apache.activemq.artemis.core.paging.cursor.NonExistentPage: Invalid messageNumber passed
While a receiving transaction is committed to a paging operation, if the page is complete, a transaction operation deletes it. Previously, the transaction operation might delete the page before the receiving transaction could access it. This caused the receiving transaction to fail to find a message. In this case, you saw an exception. This issue is now resolved.
ENTMQBR-2117 - [AMQ7,Hawtio] Unable to delete queue with single quote from Hawtio console
In prior releases, if you used the Hawtio console to delete a message containing a single quote, the broker did not successfully delete the message. This issue is now corrected.
ENTMQBR-2147 - (7.2.z) Backup doesn’t activate after shared store is reconnected
Previously, if you had a live-backup broker pair configured for high availability using shared store, activation of the backup broker upon shutdown of the live broker could fail. Specifically, this situation occurred if the shared store had previously been disconnected and reconnected, before shutdown of the live broker. This issue is now resolved.
ENTMQBR-2177 - Allow deleting temporary destination when originating session is closed
Previously, if you used the JMS API to delete a temporary destination when the originating session was already closed, the operation seemed to be successful, when it was not. This issue is now corrected.
ENTMQBR-2325 - JMSBridge fails to stop after throwing an error
Previously, a JMS bridge between JBoss EAP and a broker not using Artemis Core Protocol produced an exception while stopping if the broker was already stopped. As a result of this exception, neither the EAP server nor the JMS bridge could successfully stop. This issue is now resolved.
ENTMQBR-2326 - JMSBridge fails to stop
Previously, a JMS bridge between JBoss EAP and a broker not using Artemis Core Protocol produced an exception while stopping. As a result of this exception, neither the EAP server nor the JMS bridge could successfully stop. This issue is now resolved.
ENTMQBR-2333 - Temporary Queue Address Leak In Clustered Configuration
Previously, if a temporary queue was replicated across cluster nodes, closing connections to the temporary queue on a given node did not purge the server resources associated with the queue. In other words, the queue had an address leak. This issue is resolved. You can now use the
address-queue-scan-periodparameter to specify how often the broker scans queues for potential deletion.
ENTMQBR-2337 - Failover or similar option for AMQ Core Protocol JMS Client
In prior releases, an AMQ Core Protocol JMS Client could make repeated attempts to connect to an unavailable broker server, even if alternative broker servers were specified in the client’s
ConnectionFactoryconfiguration. Specifically, this situation occurred if you set the
-1. This type of configuration is shown in the example below.
This issue is now resolved. If the first listed server that the client tries is unavailable, the client immediately tries the next, and so on. The client repeats this loop until it establishes a broker connection.
ENTMQBR-2386 - Slow consumers not disconnected if producer msg/s rate is low
Previously, when checking for slow consumers, the broker calculated only messages that a producer added during the check period. The broker did not take into account existing messages in the queue at the start of the check period. This could cause the broker to miss legitimate cases of slow message consumption. This situation could arise, for example, if a producer previously sent large numbers of messages to the queue in bursts, but was idle during the slow consumer check period.
ENTMQBR-2464 - Qpid JMS client throws incorrect exception when sending message to activemq.management
In prior releases, when you used the AMQP Qpid JMS client to send a message to the internal queue
activemq.management, you saw an exception. This issue is now resolved.
ENTMQBR-2467 - [AMQ7, large header] message with very large header shuts broker down with "AMQ222010: Critical IO Error, shutting down the server.."
In prior releases, if a client sent a message with a header larger than the value of the
journal-buffer-sizeparameter, you saw an exception that caused the broker to shut down. This issue is resolved. The broker now rejects messages with a header larger than the value of the
ENTMQBR-2468 - java.lang.NullPointerException with message replication
When the values of the
config-delete-queuesparameters of your broker configuration are set to
force, the broker reloads the
broker.xmlconfiguration file after a change. When the broker reloads the configuration, any auto-created, clustered queues are deleted. In prior releases, this deletion caused a failure that prevented the broker from starting. This issue is now resolved. For any queues that are deleted, the broker catches an exception and logs a warning, but starts as expected.
ENTMQBR-2476 - Live server does not shutdown when using vote-on-replication-failure
Previously, in a cluster of three or more live-backup groups that was using the replication high availability (HA) policy, the live broker might not shut down when its replication connection failed. Specifically, this issue occurred when all of the following were true:
vote-on-replication-failureparameter in the HA policy configuration was set to
- The live broker failed to receive a majority vote during quorum voting
- The backup broker became live before quorum voting took place
This issue is now resolved.
ENTMQBR-2498 - Message Grouping may have messages stuck if a consumer is busy during the delivery
Previously, if you used message grouping, a message might not be delivered if the consumer was busy. In these cases, messages remained stuck on the queue until the broker was restarted. This issue is now resolved.
ENTMQBR-2526 - Clustering does not move messages to another node, when there is a divert
Previously, in a broker cluster configured with diverts, messages were not successfully redistributed from an anycast queue on one broker to an anycast queue with active consumers at the same address on another broker. Instead, messages were dropped. This issue is now resolved.
ENTMQBR-2528 - MQTT Consumer gets java.io.EOFException when subscribed with QOS 2 for an address with custom security settings
Previously, an MQTT consumer saw an exception when subscribed with quality of service (QoS) level 2 to an address with custom security settings. This issue is now resolved.
ENTMQBR-2575 - [AMQ 7.3.0] HawtIo - Can’t send Message with Headers
Previously, you could not use the Hawtio console to send messages with headers. This issue is now resolved.
ENTMQBR-2579 - Masked password does not work in LDAP security-setting-plugin
In prior releases, if you used the
artemis maskcommand to encrypt a password for inclusion in the LDAP
<security-setting-plugin>configuration, the encrypted password did not work, even if the
mask-passwordattribute in your
broker.xmlconfiguration file was set to
true. As a result of this issue, you saw an LDAP authentication error. This issue is now resolved.
ENTMQBR-2601 - Statically created addresses may be auto-deleted
Previously, the reaper thread, which deletes auto-created queues, could also delete queues that you statically defined in the
broker.xmlconfiguration file. This issue is now resolved. The reaper thread deletes only auto-created queues.
ENTMQBR-2675 - Attribute group-name ignored in replicated colocated configurations
Previously, if you configured a cluster of three master-slave broker pairs to use high availability with colocated backups, the value of the
group-nameattribute of each broker was ignored. As a result, when you started brokers in a new cluster configured this way, the first two master-slave pairs that you started immediately began replicating to each other, while the final master-slave pair that you started did not get a replication partner. The final master-slave pair then polled the cluster indefinitely, looking for a partner. This issue is now resolved.
ENTMQBR-2707 - Transaction rollback ordering is broken with AMQP
Previously, when rolling back AMQP transactional message data, the broker might fail to preserve the message order. This issue is now resolved.
ENTMQBR-2733 - JDBC HA shared store does not take credentials from the jdbc-user and jdbc-password tags
Previously, when you configured shared store high availability using JDBC persistence, the broker did not read the
jdbc-passwordvalues that you specified in the
database-storeconfiguration. To work around this issue, you could include the credentials directly in the
jdbc-Connection-URLelement. However, when specified in this way, the credentials were not encrypted. This issue is now resolved. The broker reads the encrypted values that you specify for
ENTMQBR-2876 - OpenWire Consumers are Not Cleaned up When the Session is Closed but Connection Remains Open
Previously, if an OpenWire client closed a session without closing associated consumers, the consumers were not cleaned up by the broker. Even if the client exited, an
OpenWireConnectionobject remained on the broker, with a potentially large collection in
OpenWireConnection.consumerExchanges. Over time, this situation exhausted the broker heap space and the broker entered an out-of-memory (OOM) state. This issue is now resolved.
For information about issues resolved in maintenance releases, see the following articles:
Chapter 6. Known Issues
ENTMQBR-17 - AMQ222117: Unable to start cluster connection
A broker cluster may fail to initialize properly in environments that support IPv6. The failure is due to a
SocketExceptionthat is indicated by the log message
Can’t assign requested address. To work around this issue, set the
java.net.preferIPv4Stacksystem property to
ENTMQBR-463 - Attributes in clustering settings have order restrictions. Would be nice to either have better error message or simply ignore the order
Currently the sequence of the elements in the cluster connection configuration has to be in a specific order. The workaround is to adhere to the order in the configuration schema.
ENTMQBR-520 - Receiving from address named the same as a queue bound to another address should not be allowed
A queue with the same name as an address must only be assigned to address. Creating a queue with the same name as an existing address, but bound to an address with a different name, is an invalid configuration. Doing so can result in incorrect messages being routed to the queue.
ENTMQBR-522 - Broker running on windows write problems with remove temp files when shutting down
On Windows, the broker does not successfully clean up temporary files when it shuts down. This issue causes the shutdown process to be slow. In addition, temporary files not deleted by the broker accumulate over time.
ENTMQBR-569 - Conversion of IDs from OpenWire to AMQP results in sending IDs as binary
When communicating cross-protocol from an A-MQ 6 OpenWire client to an AMQP client, additional information is encoded in the application message properties. This is benign information used internally by the broker and can be ignored.
ENTMQBR-599 - Define truststore and keystore by Artemis cli
Creating a broker instance by using the
--ssl-trust-passwordparameters does not work. To work around this issue, set the corresponding properties manually in
bootstrap.xmlafter creating the broker.
ENTMQBR-617 - Cert based authentication does not work with AMQP and OpenWire clients
Authentication based on SSL certificates is not supported for AMQP and OpenWire clients.
ENTMQBR-636 - Journal breaks, causing
JavaNullPointerException, under perf load (mpt)
To prevent IO-related issues from occurring when the broker is managing heavy loads, verify that the JVM is allocated with enough memory and heap space. See the section titled "Tuning the VM" in the Performance Tuning chapter of the ActiveMQ Artemis documentation.
ENTMQBR-648 - JMS Openwire client is unable to send messages to queue with defined
Using an A-MQ 6 JMS client to send messages to an address that has a queue with
truefails if the queue has no consumers. It is recommended that you do not set the
purgeOnNoConsumeroption when using A-MQ 6 JMS clients.
ENTMQBR-652 - List of known
This version of
amq-jon-pluginhas known issues with the MBeans for broker and queue.
Issues with the broker MBean:
Closing a connection throws
Adding address settings throws
getConnectorServices()operation cannot be found
listConsumersAsJSON()operation cannot be found
getDivertNames()operation cannot be found
Listing network topology throws
- Remove address settings has wrong parameter name
Issues with the queue MBean:
expireMessage()throws argument type mismatch exception
IllegalArgumentExceptionwith error message argument type mismatch
IllegalArgumentExceptionwith error message argument type mismatch
removeMessages()throws exception with error Can’t find operation removeMessage with 2 arguments
retryMessage()throws argument type mismatch
- Closing a connection throws
ENTMQBR-655 - [AMQP] Unable to send message when
The configuration option
populate-validated-useris not supported for messages produced using the AMQP protocol.
ENTMQBR-738 - Unable to build AMQ 7 examples offline with provided offline repo
You cannot build the examples included with AMQ Broker in an offline environment. This issue is caused by missing dependencies in the provided offline Maven repository.
ENTMQBR-897 - Openwire client/protocol issues with special characters in destination name
Currently AMQ OpenWire JMS clients cannot access queues and addresses that include the following characters in their name: comma (','), hash ('#'), greater than ('>'), and whitespace.
ENTMQBR-944 - [A-MQ7, Hawtio, RBAC] User gets no feedback if operation access was denied by RBAC
The console can indicate that an operation attempted by an unauthorized user was successful when it was not.
ENTMQBR-1498 - Diagram in management console for HA (replication, sharedstore) does not reflect real topology
If you configure a broker cluster with some extra, passive slaves, the cluster diagram in the web console does not show these passive slaves.
ENTMQBR-1815 - Hawtio view changes on automated refresh
When automatic refresh is enabled, the Hawtio console updates the screen every 5 seconds. Automatic refreshes also change the view to the Attributes screen, causing you to lose focus from any other screen you are viewing. To work around this issue, select Preferences in the top-right corner of the console. In the Update rate drop-down list, select
ENTMQBR-1848 - "javax.jms.JMSException: Incorrect Routing Type for queue, expecting: ANYCAST" occurs when qpid-jms client consumes a message from a multicast queue as
javax.jms.Queueobject with FQQN
Currently, sending a message by using the Qpid JMS client to a multicast queue by using FQQN (fully qualified queue name) to an address that has multiple queues configured generates an error message on the client, and the message cannot be sent. To work around this issue, modify the broker configuration to resolve the error and unblock the client.
ENTMQBR-1875 - [AMQ 7, ha, replicated store] backup broker appear not to go "live" or shutdown after - ActiveMQIllegalStateException errorType=ILLEGAL_STATE message=AMQ119026: Backup Server was not yet in sync with live
Removing the paging disk of a master broker while a backup broker is trying to sync with the master broker causes the master to fail. In addition, the backup broker cannot become live because it continues trying to sync with the master.
ENTMQBR-1942 - [AMQ 7.2, shared store, scale down] NullPointer exception when slave activates and tries to scale down
If you configure a slave broker to use a shared store and scaledown in a cluster, you can encounter an exception on the slave if the master broker goes offline.
ENTMQBR-2068 - some messages received but not delivered during HA fail-over, fail-back scenario
Currently, if a broker fails over to its slave while an OpenWire client is sending messages, messages being delivered to the broker when failover occurs could be lost. To work around this issue, ensure that the broker persists the messages before acknowledging them.
ENTMQBR-2452 - Upgraded broker AMQ 7.3.0 from AMQ 7.2.4 on Windows cannot log
If you intend to upgrade a broker instance from 7.2.4 to 7.3.0 on Windows, logging will not work unless you specify the correct log manager version during your upgrade process. For more information, see Upgrading from 7.2.x to 7.3.0 on Windows.
ENTMQBR-2470 - [AMQ7, openwire,redelivery] redelivery counter for message increasing, if consumer is closed without consuming any messages
If a broker sends a message to an Openwire consumer, but the consumer is closed before consuming the message, the broker wrongly increments the redelivery count for the pending message. If the number of occurrences of this behavior exceeds the value of the
max-delivery-attemptsconfiguration parameter, the broker sends the message to the dead letter queue (DLQ) or drops the message, based on your configuration. This issue does not affect other protocols, such as the Core protocol.
ENTMQBR-2593 - broker does not set message ID header on cross protocol consumption
A Qpid JMS client successfully retrieves a message ID only if the message was produced by another Qpid JMS client. If the message was produced by a Core JMS or OpenWire client, the Qpid JMS client cannot read the message ID.
ENTMQBR-2678 - After isolated master is live again it is unable to connect to the cluster
In a cluster of three or more live-backup groups that is using the replication high availability (HA) policy, the live broker shuts down when its replication connection fails. However, when the replication connection is restored and the original live broker is restarted, the broker is sometimes unable to rejoin the broker cluster. To enable the original live broker to rejoin the cluster, first stop the new live (original backup) broker, restart the original live broker, and then restart the original backup broker.
ENTMQBR-2711 - ServerSessionImpl cache does not clear names of deleted temporary destinations
When you create and delete temporary queues, the broker does not remove these queue names from the internal
ServerSessionImplcache. Over time, this data accumulates, causing a memory leak.
ENTMQBR-2767 - [AMQ7, purge message, OutOfMemoryException] with a large queue size, removeAllMessages() takes a long time and can eventually result in an OOM exception
For a queue with a large number (that is, in the order of millions) of messages, the JMX operation
removeAllMessagescan take up to several minutes to complete. In addition, the operation sometimes fails to complete, returning an out-of-memory (OOM) exception instead.
ENTMQBR-2890 - Infrequently, creating a CR instance with size n > 1 causes the the nth broker pod to start and immediately restart once
If you use a Custom Resource (CR) instance to deploy a broker cluster via the AMQ Broker Operator, the final broker Pod in the deployment (as determined by the
sizeattribute of your CR) starts and then immediately restarts one time before it is usable. To work around this issue, incrementally increase the
sizeattribute of your CR to reach the desired cluster size. First, set
1and wait for the broker Pod to start. Then, set
2, wait for the Pod to start, and so on. In this case, none of the broker Pods undergo a restart before they are usable.
ENTMQBR-2942 - Pod #0 tries to contact non-existent Pods
If you change the
sizeattribute of your Custom Resource (CR) instance to scale down a broker deployment, the first broker Pod in the cluster can make repeated attempts to connect to the drainer Pods that started up to migrate messages from the brokers that shut down, before they shut down themselves. To work around this issue, follow these steps:
1) Scale your deployment to a single broker Pod.
2) Wait for all drainer Pods to start, complete message migration, and then shut down.
3) If the single remaining broker Pod has log entries for an “unknown host exception”, scale the deployment down to zero broker Pods, and then back to one.
4) When you have verified that the single remaining broker Pod is not recording exception-based log entries, scale your deployment back to its original size.