This chapter provides the following information for each release:

  • A link to the full release notes which includes all issues resolved in the release.

  • A brief list of "highlights" when applicable.

  • If necessary, specific steps required when upgrading from the previous version.

If the upgrade spans multiple versions then the steps from each version need to be followed in order.
Follow the general upgrade procedure outlined in the Upgrading the Broker chapter in addition to any version-specific upgrade instructions outlined here.

1. 2.35.0

2. 2.34.0

2.2. Upgrading from 2.33.0

  • Due to ARTEMIS-4712 the connection pooling functionality configured via the connectionPool property in login.config is no longer supported in the LDAPLoginModule. The login.config may still use the connectionPool property. No error will be thrown. However, connections will no longer be pooled regardless of the configuration.

  • Due to ARTEMIS-4498 the web console will now report all internal objects.

    • This was done in an attempt to allow administrators to act when things are not working as expected, to get metrics on these objects and allow more transparency for the broker.

    • this includes all Openwire Advisor queues and addresses, MQTT internal objects, Cluster Store and Forward (SNF) Queues, Mirror SNF.

    • You may want to revisit authorizations if you mean to control access to certain users on the web console.

  • The CLI operation ./artemis queue stat has its output improved and updated. If you parsed the previous output in scripts you will see differences in the output.

    • It is not recommended to parse the output of a CLI Operation. You may use jolokia calls over management instead with proper JSON output.

3. 2.33.0

3.1. Highlights

  • Support for JSON formatted typed properties on CLI producer command

  • New CLI command pwd for showing directories related to the current instance

  • Maven Bill of Materials (BOM) artemis-bom to simplify integration

  • "FirstMessage" API for scheduled messages

  • New "view" and "edit" permissions for management operations configurable via security-settings in broker.xml

  • New sslAutoReload parameter for the embedded web server configured in bootstrap.xml to detect and automatically reload whe SSL stores change on disk

  • Performance improvements on mirroring and paging

  • Logging metrics to mitigate the risk of missing WARN or ERROR messages in the log.

  • Much improved documentation on network isolation (aka split brain)

  • Pluggable lock manager (aka pluggable quorum voting) out of "experimental" status and ready for general use

3.2. Upgrading from 2.32.0

  • Due to ARTEMIS-4532 the names of addresses and queues related to MQTT topics and subscriptions respectively may change. This will impact MQTT use-cases if both of the following are true:

    1. The broker is configured to use a wildcard syntax which doesn’t match the MQTT wildcard syntax (e.g. the default wildcard syntax).

    2. You are using characters from the broker’s wildcard syntax in your MQTT topic name or filter. For example, if you were using the default wildcard syntax and an MQTT topic named 1.0/group/device. The dot (.) character here is part of the broker’s wildcard syntax, and it is being used in the name of an MQTT topic.

      In this case the characters from the broker’s wildcard syntax that do not match the characters in the MQTT wildcard syntax will be escaped with a backslash (i.e. \). To avoid this conversion you can configure the broker to use the MQTT wildcard syntax or change the name of the MQTT topic name or filter.

    This change will also impact OpenWire JMS consumers which are using # instead of < for wildcard purposes. In previous versions the # character was just passed through when converting from the OpenWire wildcard format to the Core wildcard format. However, now the # character is escaped during conversion. It is a bug for an application to use to use # as a wildcard with the OpenWire JMS client; > is the proper character to use as specified in the ActiveMQ Classic documentation on wildcards.

  • Due to ARTEMIS-4559 folks embedding the broker and also depending on the artemis-quorum-ri and/or artemis-quorum-api modules and/or using org.apache.activemq.artemis.core.config.ha.DistributedPrimitiveManagerConfiguration will need to use artemis-lockmanager-ri, artemis-lockmanager-api, and org.apache.activemq.artemis.core.config.ha.DistributedLockManagerConfiguration respectively. Previously these were marked as "experimental" in the documentation and were changed strictly in name to clarify their use conceptually. Furthermore, the documentation around high availability and network isolation (i.e. split brain) was refactored significantly to be more clear and comprehensive.

4. 2.32.0

4.1. Highlights

4.2. Upgrading from 2.31.x

  • Due to ARTEMIS-4501 MQTT subscription queues will be automatically removed when the corresponding session expires, either based on the session expiry interval passed by an MQTT 5 client or based on the configured defaultMqttSessionExpiryInterval for MQTT 3.x clients or MQTT 5 clients which don’t explicitly pass a session expiry interval.

    Prior to this change removing subscription queues relied on the generic auto-delete-* address-settings.

    These settings are now no longer required.

    Configure defaultMqttSessionExpiryInterval instead.

  • Due to ARTEMIS-3474 the following configuration elements have changed wherever they occur (e.g. broker.xml, bootstrap.xml, etc.), although all the previous configurations will still be supported for the time being:

    • masterprimary

    • slavebackup

    • check-for-live-servercheck-for-active-server

    • whitelistallowlist

    • blacklistdenylist

    Additionally, references to these elements have also changed in the documentation and in management interfaces. Cluster topology information (e.g. returned from the listNetworkTopology) will contain both primary and live entries for nodes functioning as primary servers.

5. 2.31.2

6. 2.31.1

6.1. Highlights

  • Bug Fixes and component upgrades

7. 2.31.0

7.1. Highlights

  • Introduced an interactive shell for running CLI command as well as Bash & ZSH auto-complete support.

  • Added a CLI cluster verification tool to help monitor broker topologies. Use via the check cluster command.

  • The queue stat command is now able to to verify the message counts on the entire cluster topology when clustering is in use.

  • Added AMQP Federation support to broker connections.

  • MQTT subscription state is now persisted.

  • Significantly improved the Paging JDBC Persistence.

  • Converted much of the documentation from MarkDown to AsciiDoc. See ARTEMIS-4383 for more details.

  • Many other bug fixes and improvements.

7.2. Upgrading from 2.30.0

  • Due to ARTEMIS-4372 and the introduction of the new Artemis shell feature when you invoke ./artemis it will now start the new shell to navigate through the CLI commands rather than just spitting out the help text.

8. 2.30.0

8.1. Highlights

  • This is mainly a bug-fix release with a few small improvements and a handful of dependency upgrades. See the release notes for all the details.

9. 2.29.0

9.1. Highlights

  • This version underwent extensive testing and fixes regarding Large Messages, with a few JIRAs dedicated to this topic. Look on the release notes for more information.

9.2. Upgrading from 2.28.0

  • Due to ARTEMIS-4151 the default access for MBeans not defined in the role-access or allowlist of management.xml is now read only. This is a precautionary measure to ensure no unanticipated MBean deployed with the broker poses a risk. However, this will also impact JVM-specific and platform MBeans as well (e.g. which allow manual garbage collection, "flight recording," etc.). Write access and general operational access to these MBeans will now have to be manually enabled in management.xml either by changing the default-access (not recommended) or specifically configuring a role-access for the particular MBean in question.

    This applies to all MBean access including directly via JMX and via the Jolokia JMX-HTTP bridge.
  • Due to ARTEMIS-4212 the broker will reject address definitions in broker.xml which don’t specify a routing type, e.g.:

    <address name="myAddress"/>

    Such configurations will need to be changed to specify a routing-type, e.g.:

    <address name="myAddress">


    <address name="myAddress">

    If an address without a routing type is configured the broker will throw an exception like this and fail to start:

    java.lang.IllegalArgumentException: AMQ229247: Invalid address configuration for 'myAddress'. Address must support multicast and/or anycast.
          at org.apache.activemq.artemis.core.deployers.impl.FileConfigurationParser.parseAddressConfiguration(
          at org.apache.activemq.artemis.core.deployers.impl.FileConfigurationParser.parseAddresses(
          at org.apache.activemq.artemis.core.deployers.impl.FileConfigurationParser.parseMainConfig(
          at org.apache.activemq.artemis.core.config.impl.FileConfiguration.parse(
          at org.apache.activemq.artemis.core.config.FileDeploymentManager.readConfiguration(
          at org.apache.activemq.artemis.integration.FileBroker.createComponents(
          at org.apache.activemq.artemis.cli.commands.Run.execute(
          at org.apache.activemq.artemis.cli.Artemis.internalExecute(
          at org.apache.activemq.artemis.cli.Artemis.execute(
          at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
          at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(
          at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(
          at java.base/java.lang.reflect.Method.invoke(
          at org.apache.activemq.artemis.boot.Artemis.execute(
          at org.apache.activemq.artemis.boot.Artemis.main(
  • Due to ARTEMIS-3707 all use of javax.transaction.TransactionManager was removed from the JCA Resource Adapter. However, this rendered the transactionTimeout activation configuration property useless. Some existing users rely on this behavior so it has been restored and properly deprecated for future removal.

10. 2.28.0

10.1. Highlights

  • Bug Fixes and improvements as usual

  • ARTEMIS-4136 Mirror sync replication

    • Mirror now has an option to set sync=true. Blocking operations from clients will wait a round trip on the mirror.

  • ARTEMIS-4065 Paging Counter Journal Records were removed

    • We don’t store page counters records on the journal any longer what should simplify operation and improve performance.

10.2. Upgrading from 2.27.0

  • Due to ARTEMIS-3871 the naming pattern used for MQTT shared subscription queues has changed. Previously the subscription queue was named according to the subscription name provided in the MQTT SUBSCRIBE packet. However, MQTT allows the same name to be used across multiple subscriptions whereas queues in the broker must be named uniquely. Now the subscription queue will be named according to the subscription name and topic name so that all subscription queue names will be unique. Before upgrading please ensure all MQTT shared subscriptions are empty. When the subscribers reconnect they will get a new subscription queue. If they are not empty you can move the messages to the new subscription queue administratively.

11. 2.27.1

11.1. Highlights

  • Bug Fixes

  • AMQP Large Message over Bridges were broken

  • Rollback of massive transactions would take a long time to process

  • Improvements to auto-create and auto-delete queues.

12. 2.27.0

12.1. Highlights

  • 2.27.0 Introduced a new upgrade tool to help migrating your instance to a newer version.

  • The client and broker now use SLF4J for their logging API.

  • The broker distribution now uses Log4J 2 as its logging implementation.

12.2. Upgrading from 2.26.0

Client applications wanting logging will now need to supply an appropriate SLF4J-supporting logging implementation configured appropriately for their needs. See client application logging for more information plus an example around using Log4J 2.

The broker distribution now includes and configures Log4J 2 as its logging implementation, see logging for more details. If upgrading an existing broker instance rather than creating a new instance, some configuration etc updates will be necessary for the brokers existing instance /etc and /bin files.

You can use the new upgrade helper tool from the newly downloaded broker to refresh various configuration files and scripts for an existing broker instance. The broker.xml and data are left in place as-is.

You should back up your existing broker instance before running the command.

The command can be executed by running ./artemis upgrade <path-to-your-instance> from the new downloaded broker home.

Most existing customisations to the old configuration files and scripts will be lost in the process of refreshing the files. As such you should compare the old configuration files with the refreshed ones and then port any missing customisations you may have made as necessary. The upgrade command itself will copy the older files it changes to an old-config-bkp. folder within the instance directory.

Similarly, if you had customised the old file you may need to prepare analogous changes for the new file.

Note also that the configuration-file-refresh-period setting in broker.xml no longer covers logging configuration refresh. Log4J 2 has its own configuration reload handling, configured via the monitorInterval property within the Log4J configuration file itself. The default <instance>/etc/ file created has a 5 second monitorInterval value set to align with the prior default broker behaviour.

12.3. Manual update

Alternatively, rather than using the upgrade helper command as outlined above, you can instead perform the update manually, following the general upgrading procedure plus the additional steps below:

  1. The new <instance>/etc/ file should be created with Log4J 2 configuration. The file used by the "artemis create" CLI command can be downloaded from:

  2. The old <instance>/etc/ JBoss Logging configuration file should be deleted.

  3. Related startup script or profile cleanups are needed: a diff file demonstrating the changes needed since 2.26.0 is available here for *nix or here for Windows.

13. 2.26.0

13.1. Highlights

  • Bug fixes and improvements

13.2. Upgrading from 2.25.0

  1. Due to ARTEMIS-4006 the artemis-jms-client-all and artemis-jakarta-client-all clients were removed from the lib/client directory in the binary distribution. If you use these libraries they can be found at Maven Central (e.g. here). Please refer to the client class path documentation for more information.

  2. We removed the REST interface from the code-base and documentation. If you still require the REST interface you can access the latest version which is still viable. You can still follow the steps from the previous documentation to build and deploy the interface. However, you should stop using it as it will not be maintained any more.

  3. Due to ARTEMIS-3980 the web content was removed from the binary distribution. We now redirect web requests with the root target to the administration console. To enable this new redirect behavior on current instances you have to update bootstrap.xml. Change:

    <web path="web">


    <web path="web" rootRedirectLocation="console">

    If you used to customize the index page or to add custom content in the web folder please refer to the web-server documentation for more information on disabling the redirect and enabling the web content.

14. 2.25.0

14.1. Highlights

  • Improvement on Paging Flow Control

  • Many other bug fixes and improvements

15. 2.24.0

15.1. Highlights

  • Streamlined page caches and files are just read into queues without the need of soft caches.

15.2. Upgrading from 2.23.0

  1. Due to ARTEMIS-3851 the queue created for an MQTT 3.x subscriber using CleanSession=1 is now non-durable rather than durable. This may impact security-settings for MQTT clients which previously only had createDurableQueue for their role. They will now need createNonDurableQueue as well. Again, this only has potential impact for MQTT 3.x clients using CleanSession=1.

  2. Due to ARTEMIS-3892 the username assigned to queues will be based on the validated user rather than just the username submitted by the client application. This will impact use-cases like the following:

    1. When login.config is configured with the GuestLoginModule which causes some users to be assigned a specific username and role during the authentication process.

    2. When login.config is configured with the CertificateLoginModule which causes users to be assigned a username and role corresponding to the subject DN from their SSL certificate.

    In these kinds of situations the broker will use this assigned (i.e. validated) username for any queues created with the connection. In the past the queue’s username would have been left blank.

16. 2.23.1

16.1. Highlights

  • ARTEMIS-3856 - Failed to change channel state to ReadyForWriting : java.util.ConcurrentModificationException

17. 2.23.0

18. 2.22.0

18.1. Highlights

  • The default producer-window-size on cluster-connection was changed to 1MB to mitigate potential OutOfMemoryErrors in environments with with high latency networking.

19. 2.21.0

19.1. Highlights

19.2. Upgrading from 2.20.0

  1. Due to XML schema changes to correct an inaccurate domain name 2 files will need to be updated:

    1. etc/bootstrap.xml

    2. etc/management.xml

      In both files change the XML namespace from to, e.g. in bootsrap.xml use:

       <broker xmlns="">

      And in management.xml use:

       <management-context xmlns="">
  2. If you’re using JDBC persistence then due to the changes in ARTEMIS-3679 you’ll need to update your database. The column HOLDER_EXPIRATION_TIME on the NODE_MANAGER_STORE changed from a TIMESTAMP to a BIGINT (or NUMBER(19) on Oracle). You will have to stop any broker that is accessing that table and either drop it or execute the proper ALTER TABLE statement for your database. If you drop the table then it will be automatically recreated when broker restarts and repopulated with a new, auto-generated node ID.

  3. If you’re using JGroups then due to the changes in ARTEMIS-2413 where JGroups was updated from 3.x to 5.x you will need to update your JGroups configuration. Many of the protocols have changed, and there’s no automated tool to bring legacy configurations up to date so please refer to the JGroups documentation for more details on the new configuration. You can find example configurations in the JGroups repository (e.g. tcp.xml and udp.xml).

20. 2.20.0

20.1. Highlights

  • Java 11 is now required.

21. 2.19.0

21.1. Highlights

  • New ability to replay retained journal records via the management API.

  • New environment/system property to set the "key" for masked passwords when using the default codec.

  • Ability to disable message-load-balancing and still allow redistribution via the new OFF_WITH_REDISTRIBUTION type.

  • MQTT session state can now be cleaned up automatically to avoid excessive accumulation in situations where client’s don’t clean up their own sessions.

  • Distribute full Jakarta Messaging 3.0 client in the lib/client directory along with a new example of how to use it in examples/features/standard/queue-jakarta.

22. 2.18.0

22.1. Highlights

22.2. Upgrading from 2.17.0

  1. Due to ARTEMIS-3367 the default setting for verifyHost on core connectors has been changed from false to true. This means that core clients will now expect the CN or Subject Alternative Name values of the broker’s SSL certificate to match the hostname in the client’s URL.

    This impacts all core-based clients including core JMS clients and core connections between cluster nodes. Although this is a "breaking" change, not performing hostname verification is a security risk (e.g. due to man-in-the-middle attacks). Enabling it by default aligns core client behavior with industry standards. To deal with this you can do one of the following:

    • Update your SSL certificates to use a hostname which matches the hostname in the client’s URL. This is the recommended option with regard to security.

    • Update any connector using sslEnabled=true to also use verifyHost=false. Using this option means that you won’t get the extra security of hostname verification, but no certificates will need to change. This essentially restores the previous default behavior.

    For additional details about please refer to section 3.1 of RFC 2818 "HTTP over TLS".

  2. Due to ARTEMIS-3117 SSL keystore and truststores are no longer reloaded automatically. Previously an instance of was created for every connection. This would implicitly pick up any changes to the keystore and truststore for any new connection. However, this was grossly inefficient and therefore didn’t scale well with lots of connections. The behavior was changed so that just one is created for each acceptor. However, one can still reload keystores & truststores from disk without restarting the broker. Simply use the reload management operation on the acceptor. This is available via JMX, the web console, Jolokia, etc.

    Here’s an example curl command you can use with Jolokia to invoke the artemis acceptor’s reload operation:

    curl --user admin:admin --header "Content-Type: application/json" --request POST --data '{"type":"exec", "mbean":"org.apache.activemq.artemis:broker=\"\",component=acceptors,name=\"artemis\"", "operation":"reload"}' http://localhost:8161/console/jolokia/exec

    Of course you’ll want to adjust the username & password as well as the broker and acceptor names for your environment.

  3. The "rate" metric for queues was removed from the web console via ARTEMIS-3397. This was a follow-up from ARTEMIS-2909 in 2.16.0 (referenced in the upgrade instructions below). The "rate" metric mistakenly left visible on the web console after it was removed from the management API.

  4. Due to ARTEMIS-3141, ARTEMIS-3128, & ARTEMIS-3175 the data returned for any "list" or "browse" management method which return message data, including those exposed via the web console, will have their return data truncated by default. This is done to avoid adverse conditions with large volumes of message data which could potentially negatively impact broker stability. The management-message-attribute-size-limit address-setting controls this behavior. If you wish to restore the previous (and potentially dangerous behavior) then you can specify -1 for this. It is 256 by default.

23. 2.17.0

23.1. Highlights

  • Message-level authorization similar to ActiveMQ "Classic".

  • A count of addresses and queues is now available from the management API.

  • You can now reload the broker’s configuration from disk via the management API rather than waiting for the periodic disk scan to pick it up

  • Performance improvements on libaio journal.

  • New command-line option to transfer messages.

  • Performance improvements for the wildcard address manager.

  • JDBC datasource property values can now be masked.

  • Lots of usability improvements to the Hawtio 2 based web console introduced in 2.16.0

  • New management method to create a core bridge using JSON-based configuration input.

  • Jakarta Messaging 2.0 & 3.0 artifacts for Jakarta EE 8 & 9 respectively.

24. 2.16.0

24.1. Highlights

  • Configurable namespace for temporary queues

  • AMQP Server Connectivity

  • "Basic" SecurityManager implementation that supports replication

  • Consumer window size support for individual STOMP clients

  • Improved JDBC connection management

  • New web console based on Hawtio 2

  • Performance optimizations (i.e. caching) for authentication and authorization

  • Support for admin objects in the JCA resource adapter to facilitate deployment into 3rd-party Java EE application servers

  • Ability to prevent an acceptor from automatically starting

24.2. Upgrading from 2.15.0

  1. Due to ARTEMIS-2893 the fundamental way user management was implemented had to change to avoid data integrity issues related to concurrent modification. From a user’s perspective two main things changed:

    1. User management is no longer possible using the artemis user commands when the broker is offline. Of course users are still free to modify the properties files directly in this situation.

    2. The parameters of the artemis user commands changed. Instead of using something like this:

      ./artemis user add --user guest --password guest --role admin

      Use this instead:

      ./artemis user add --user-command-user guest --user-command-password guest --role admin

      In short, use user-command-user in lieu of user and user-command-password in lieu of password. Both user and password parameters now apply to the connection used to send the command to the broker.

      For additional details see ARTEMIS-2893 and ARTEMIS-3010

  2. Due to ARTEMIS-2909 the "rate" metric was removed from the management API for queues. In short, the org.apache.activemq.artemis.core.server.Queue#getRate method is for slow-consumer detection and is designed for internal use only.

    Furthermore, it’s too opaque to be trusted by a remote user as it only returns the number of message added to the queue since the last time it was called. The problem here is that the user calling it doesn’t know when it was invoked last. Therefore, they could be getting the rate of messages added for the last 5 minutes or the last 5 milliseconds. This can lead to inconsistent and misleading results.

    There are three main ways for users to track rates of message production and consumption (in recommended order):

    1. Use a metrics plugin. This is the most feature-rich and flexible way to track broker metrics, although it requires tools (e.g. Prometheus) to store the metrics and display them (e.g. Grafana).

    2. Invoke the getMessageCount() and getMessagesAdded() management methods and store the returned values along with the time they were retrieved. A time-series database is a great tool for this job. This is exactly what tools like Prometheus do. That data can then be used to create informative graphs, etc. using tools like Grafana. Of course, one can skip all the tools and just do some simple math to calculate rates based on the last time the counts were retrieved.

    3. Use the broker’s message counters. Message counters are the broker’s simple way of providing historical information about the queue. They provide similar results to the previous solutions, but with less flexibility since they only track data while the broker is up and there’s not really any good options for graphing.

25. 2.15.0

25.1. Highlights

  • Ability to use FQQN syntax for both security-settings and JNDI lookup

  • Support pausing dispatch during group rebalance (to avoid potential out-of-order consumption)

  • Socks5h support

26. 2.14.0

26.1. Highlights

  • Management methods to update diverts

  • Ability to "disable" a queue so that messages are not routed to it

  • Support JVM GC & thread metrics

  • Support for resetting queue properties by unsetting them in broker.xml

  • Undeploy diverts by removing them from broker.xml

  • Add addressMemoryUsagePercentage and addressSize as metrics

26.2. Upgrading from 2.13.0

This is likely a rare situation, but it’s worth mentioning here anyway. Prior to 2.14.0 if you configured a parameter on a queue in broker.xml (e.g. max-consumers) and then later removed that setting the configured value you set would remain. This has changed in 2.14.0 via ARTEMIS-2797. Any value that is not explicitly set in broker.xml will be set back to either the static default or the dynamic default configured in the address-settings (e.g. via default-max-consumers in this example). Therefore, ensure any existing queues have all the needed parameters set in broker.xml values before upgrading.

27. 2.13.0

27.1. Highlights

27.2. Upgrading from 2.12.0

Version 2.13.0 added new audit logging which is logged at INFO level and can be very verbose. The shipped with this new version is set up to filter this out by default. If your isn’t updated appropriately this audit logging will likely appear in your console and artemis.log file assuming you’re using a logging configuration close to the default. Add this to your

# to enable audit change the level to INFO


#Audit logger

formatter.AUDIT_PATTERN.pattern=%d [AUDIT](%t) %s%E%n

28. 2.12.0

28.1. Highlights

  • Support for SOCKS proxy

  • Real large message support for AMQP

  • Automatic creation of dead-letter resources akin to ActiveMQ 5’s individual dead-letter strategy

  • Automatic creation of expiry resources

  • Improved API for queue creation

  • Allow users to override JAVA_ARGS via environment variable

  • Reduce heap usage during journal loading during broker start-up

  • Allow server header in STOMP CONNECTED frame to be disabled

  • Support disk store used percentage as an exportable metric (e.g. to be monitored by tools like Prometheus, etc.)

  • Ability to configure a "customizer" for the embedded web server

  • Improved logging for errors when starting an acceptor to more easily identify the acceptor which has the problem.

  • The CLI will now read the broker.xml to find the default connector URL for commands which require it (e.g. consumer, producer, etc.)

29. 2.11.0

29.1. Highlights

30. 2.10.0

This was mainly a bug-fix release with a notable dependency change impacting version upgrade.

30.1. Upgrading from 2.9.0

Due to the WildFly dependency upgrade the broker start scripts/configuration need to be adjusted after upgrading.

30.1.1. On *nix

Locate this statement in bin/artemis:


This needs to be replaced with this:


30.1.2. On Windows

Locate this part of JAVA_ARGS in etc/artemis.profile.cmd respectively bin/artemis-service.xml:


This needs to be replaced with this:


31. 2.9.0

This was a light release. It included a handful of bug fixes, a few improvements, and one major new feature.

32. 2.8.1

This was mainly a bug-fix release with a notable dependency change impacting version upgrade.

32.1. Upgrading from 2.8.0

Due to the dependency upgrade made on ARTEMIS-2319 the broker start scripts need to be adjusted after upgrading.

32.1.1. On *nix

Locate this if statement in bin/artemis:

if [ -z "$LOG_MANAGER" ] ; then
 # this is the one found when the server was created

This needs to be replaced with this block:

if [ -z "$LOG_MANAGER" ] ; then
 # this is the one found when the server was created

WILDFLY_COMMON=`ls $ARTEMIS_HOME/lib/wildfly-common*jar 2>/dev/null`
if [ -z "$WILDFLY_COMMON" ] ; then
 # this is the one found when the server was created

Notice that the jboss-logmanager version has changed and there is also a new wildfly-common library.

Not much further down there is this line:

-Xbootclasspath/a:"$LOG_MANAGER" \

This line should be changed to be:

-Xbootclasspath/a:"$LOG_MANAGER:$WILDFLY_COMMON" \

32.1.2. On Windows

Locate this part of JAVA_ARGS in etc/artemis.profile.cmd respectively bin/artemis-service.xml:


This needs to be replaced with this:


33. 2.8.0

33.1. Highlights

34. 2.7.0

34.1. Highlights

35. 2.6.4

This was mainly a bug-fix release with a few improvements a couple notable new features:

35.1. Highlights

  • Added the ability to set the text message content on the producer CLI command.

  • Support reload logging configuration at runtime.

36. 2.6.3

This was mainly a bug-fix release with a few improvements but no substantial new features.

37. 2.6.2

This was a bug-fix release with no substantial new features or improvements.

38. 2.6.1

This was a bug-fix release with no substantial new features or improvements.

39. 2.6.0

39.1. Highlights

40. 2.5.0

40.1. Highlights

40.2. Upgrading from 2.4.0

  1. Due to changes from ARTEMIS-1644 any acceptor that needs to be compatible with HornetQ and/or Artemis 1.x clients needs to have anycastPrefix=jms.queue.;multicastPrefix=jms.topic. in the acceptor url. This prefix used to be configured automatically behind the scenes when the broker detected these old types of clients, but that broke certain use-cases with no possible work-around. See ARTEMIS-1644 for more details.

41. 2.4.0

41.1. Highlights

41.2. Upgrading from 2.3.0

  1. Create <ARTEMIS_INSTANCE>/etc/management.xml. At the very least, the file must contain this:

    <management-context xmlns=""/>

    This configures role based authorisation for JMX. Read more in the Management documentation.

  2. If configured, remove the Jolokia war file from the web element in <ARTEMIS_INSTANCE>/etc/bootstrap.xml:

    <app url="jolokia" war="jolokia.war"/>

    This is no longer required as the Jolokia REST interface is now integrated into the console web application.

    If the following is absent and you desire to deploy the web console then add:

    <app url="console" war="console.war"/>
    the Jolokia REST interface URL will now be at http://<host>:<port>/console/jolokia

42. 2.3.0

42.1. Highlights

42.2. Upgrading from 2.2.0

  1. If you desire to deploy the web console then add the following to the web element in <ARTEMIS_INSTANCE>/etc/bootstrap.xml:

    <app url="console" war="console.war"/>

43. 2.2.0

43.1. Highlights


43.2. Highlights

44. 2.0.0

44.1. Highlights

  • Huge update involving a significant refactoring of the addressing model yielding the following benefits:

    • Simpler and more flexible XML configuration.

    • Support for additional messaging use-cases.

    • Eliminates confusing JMS-specific queue naming conventions (i.e. "jms.queue." & "jms.topic." prefixes).

  • Pure encoding of messages so protocols like AMQP don’t need to convert messages to "core" format unless absolutely necessary.

  • "MAPPED" journal type for increased performance in certain use-cases.

45. 1.5.6

46. 1.5.5

47. 1.5.4

47.1. Highlights

  • Support Oracle12C for JDBC persistence.

  • Bug fixes.

48. 1.5.3

48.1. Highlights

  • Support "byte notation" (e.g. "K", "KB", "Gb", etc.) in broker XML configuration.

  • CLI command to recalculate disk sync times.

  • Bug fixes.

49. 1.5.2

49.1. Highlights

  • Support for paging using JDBC.

  • Bug fixes.

50. 1.5.1

50.1. Highlights

  • Support outgoing connections for AMQP.

  • Bug fixes.

51. 1.5.0

51.1. Highlights

  • AMQP performance improvements.

  • JUnit rule implementation so messaging resources like brokers can be easily configured in tests.

  • Basic CDI integration.

  • Store user’s password in hash form by default.

52. 1.4.0

52.1. Highlights

  • "Global" limit for disk usage.

  • Detect and reload certain XML configuration changes at runtime.

  • MQTT interceptors.

  • Support adding/deleting queues via CLI.

  • New "browse" security permission for clients who only wish to look at messages.

  • Option to populate JMSXUserID.

  • "Dual authentication" support to authenticate SSL-based and non-SSL-based clients differently.

53. 1.3.0

53.1. Highlights

  • Better support of OpenWire features (e.g. reconnect, producer flow-control, optimized acknowledgements)

  • SSL keystore reload at runtime.

  • Initial support for JDBC persistence.

  • Support scheduled messages on last-value queue.

54. 1.2.0

54.1. Highlights

  • Improvements around performance

  • OSGi support.

  • Support functionality equivalent to all 5.x JAAS login modules including:

    • Properties file

    • LDAP

    • SSL certificate

    • "Guest"

55. 1.1.0

55.1. Highlights

  • MQTT support.

  • The examples now use the CLI programmatically to create, start, stop, etc. servers reflecting real cases used in production.

  • CLI improvements. There are new tools to compact the journal and additional improvements to the user experience.

  • Configurable resource limits.

  • Ability to disable server-side message load-balancing.

56. 1.0.0

56.1. Highlights

  • First release of the donated code-base as ActiveMQ Artemis!

  • Lots of features for parity with ActiveMQ "Classic" including:

    • OpenWire support

    • AMQP 1.0 support

    • URL based connections

    • Auto-create addresses/queues

    • Jolokia integration