WSO2 Message Broker 3.2.0 BETA

WSO2 MB Team is pleased to announce the release of WSO2 Message Broker 3.2.0-BETA. Following are the bug fixes, improvements, and tasks that have been fixed along with this release.

Bug

• [MB-1645] - Durable topic subscription create queues in carbonsuper domain when the connection URL contains "clientID"
• [MB-1646] - There is a possible race condition in SlotDeliveryWorker Class
• [MB-1663] - Fix MB Issues with Multiple User stores
• [MB-1671] - SlotDeliveryWorker is not executing the assigned tasks
• [MB-1703] - Running JmsSelector Sample result warning in the wso2carbon log
• [MB-1711] - Getting a NullPointer when a consumer is closed without closing session
• [MB-1715] - Configuration issue in mqtt oauth
• [MB-1723] - Node ID is not set into Hazelcast map before "Member added" event arrives to a node.
• [MB-1727] - [Durable Subscription] When the db connectivity is dropped and reconnected durable topic subscription should be displayed as inactive
• [MB-1729] - Forbidden error when browsing the last message published via MC
• [MB-1732] - Disable queue subscription 'close' operation for non-owned nodes
• [MB-1734] - Error observed when messages moved to DLC
• [MB-1737] - With the current implementation of the moving messages from DLC can cause for an OOM issue.
• [MB-1738] - [Queue-Selectors] Some messages move to DLC, when multiple subscribers connected to the same queue via different cluster nodes
• [MB-1743] - Messages getting stuck after Forced restart
• [MB-1747] - Some messages getting stuck in the queue when the subscription owned node forcefully shutdown
• [MB-1750] - Null pointer exception during start up, when there are messages in DLC
• [MB-1753] - Messages re-routed/restored before its expiration time did not removed after the defined deletion time
• [MB-1754] - Expired messages when "expireMessagesInDLC = false" not getting delete when "expireMessagesInDLC = true"
• [MB-1755] - Expired messages not deleted from the queue
• [MB-1758] - Unable to close the queue subscriptions in a tenant domain
• [MB-1767] - [Durable Subscriptions] Expired messages in the DLC not deleted after defined time interval
• [MB-1769] - [Topics / Failover] Publisher class is unable to publish messages when a new coordinator is elected in MB cluster
• [MB-1770] - [Cluster / Coordinator] When stopping the coordinator node while messages are still publishing, a Hazelcast error is displayed
• [MB-1776] - Active durable subscriptions appeared under 'Inactive - durable subscription' list
• [MB-1780] - Queue/Topic subscriptions not enabled with the relevant permissions
• [MB-1781] - [Start up] A Nullpointer Exception occurred during the start up of MB node2
• [MB-1783] - [Tenant Domain] Un-acknowledged messages for durable topics appeared in the DLC of the carbon super user
• [MB-1786] - Nullpointer Exception observed when creating durable subscription with an existing subscription ID
• [MB-1792] - Queue Subscriptions failed when adding multiple queue subscribers to all cluster nodes
• [MB-1793] - [Unacked Messages] maxUnackedMessages value should not be hardcoded in backend
• [MB-1794] - [Intermittent] Nullpointer exception observed when queue subscriptions synced through cluster nodes
• [MB-1795] - [IndexOutOfBoundsException] Exception occurred while publishing and subscribing to a queue concurrently
• [MB-1797] - Exception in JMS client when retrieving the message count
• [MB-1798] - 'MixedDurableTopicTestCase' fails with an assert failure
• [MB-1800] - [Tenant / Durable Subscription] When creating a durable subscription for a tenant domain, permission denied error is displayed
• [MB-1801] - Created Topics subscribers missing in the subscriptions list
• [MB-1804] - [Hazelcast] Hazelcast error when stopping the node after starting with -Dsetup
• [MB-1806] - [RDBMS based coordination] Cordinator left the cluster throwing - Error detected while running coordination algorithm continuously
• [MB-1807] - [Cluster / Long running] One node has gone OOM in MB cluster while running publisher/subscriber for both Topic and Queues
• [MB-1808] - RDBMS connection is not closed in some methods
• [MB-1809] - Active durable topic subscribers are shown as active and inactive in management console
• [MB-1810] - [Cluster / Long running] Queues are not created at first in one node of the 3 node cluster
• [MB-1811] - [sporadic] All JobPoolThreads getting stuck after several AMQConnectionExceptions
• [MB-1814] - NullPointer Exception when shutting down the MB servers
• [MB-1815] - Although no active subscribers available topic cannot be deleted registry shows a entry of a subcription
• [MB-1816] - [Cluster / Long running] Some of the subscribers do not get created when execute the mixed subscriber script for Node1 Node2 in a clustered set up
• [MB-1819] - Refresh button failing for shared durable subscriptions
• [MB-1821] - Provided Inactive durable subscriber ID fails to retrieve the subscriber
• [MB-1822] - Issue when closing subscriber on failover
• [MB-1823] - Exception thrown when adding subscriptions
• [MB-1825] - [Durable Topics] Unable to search messages in DLC by its 'Destination' name
• [MB-1829] - NPE observed when restore/reroute expired messages from the DLC
• [MB-1832] - [GC Logs] GC logs are not created in repository/logs folder
• [MB-1835] - Exception thrown when syncing message routers in an RDBMS based cluster
• [MB-1840] - [Intermittent] ClassCastException when filtering durable subscription by ID
• [MB-1841] - Observed the below exception when shutting down the server
• [MB-1843] - [rdbmsBasedCoordination] On server startup MB throws - Internal error processing updateCurrentMessageIdForSafeZone
• [MB-1844] - Nullpointer thrown at startup and server stops if rdbmsBasedCoordination and networkPartitionsDetection is enabled
• [MB-1845] - Messages are remaining in DB if non durable topic subscriber is stopped while messages are published
• [MB-1850] - Messages published after a subscriber was closed and unsubscribed are delivered to the subscriber if resubscribed
• [MB-1855] - Exception thrown when adding subscriptions concurrently
• [MB-1856] - If user with no permission tries to subscribe as a durable subscription it creates an inactive durable subscription
• [MB-1857] - [Intermittent] Shared durable topic subscriber shown as active when it is in inactive state
• [MB-1858] - When durable topic subscription is closed from management console client is not notified
• [MB-1859] - Invalid channel message state transition from CLOSED suggested: NACKED printed when multiple queue subscribers subscribed to a queue
• [MB-1860] - Active non durable topic subscriber fails to consume messages
• [MB-1862] - Error at startup when netoworkPartitionDetection is enabled and does not start the server.
• [MB-1863] - Bindings are duplicated in database when there are multiple topic subscribers
• [MB-1869] - Thrift client blocks parallel calls due to synchronized behaviour
• [MB-1873] - Error when creating and closing shared subscribers between two nodes
• [MB-1874] - Queue subscription list page does not show all subscriptions by default
• [MB-1876] - Fix failing multipleSubsWithSameIdTest
• [MB-1880] - Thrift exception occurrd in RDBMSCoordinationStrategy elected coordinator
• [MB-1885] - Exception thrown when subscription pages are viewed in oracle
• [MB-1886] - Exception thrown when adding shared subscriptions concurrently to 2 nodes
• [MB-1887] - In client acknowledgement mode, when calling session.recover() continuously, messages are lost
• [MB-1888] - Sync queue purge deletes messsages from the store - leads message loss in RDBMS coordination mode
• [MB-1889] - Durable topic subscribers do not get messages even if they are active
• [MB-1890] - Add Message store error handling for Dtx DB writes
• [MB-1891] - durable subscription id is appended to subscription id in topic subscription UI
• [MB-1892] - Nullpointer exception while topic mix subscription
• [MB-1893] - Invalid message transition from EXPIRED >> SLOT_REMOVED
• [MB-1897] - Inconsistent behaviour when handling topic subscribe/publish permissions
• [MB-1899] - "Temporary Topic subscriptions" should be renamed Topic subscription list page
• [MB-1900] - Issue with subscriptions >> viewTopicSubscriptions permission
• [MB-1901] - When content read fails messages were deleted from store.
• [MB-1902] - Errors occur when queues are created and subscribed to from different cases
• [MB-1903] - Permission issue with topics
• [MB-1904] - Failure of adding a single binding/subscription when synced with db, prevents all from being added
• [MB-1906] - NPE thrown when failing-over subscribers
• [MB-1908] - An empty slot gets created in the database when no publishers are connected
• [MB-1909] - Safe zone stops getting updated in a fail-over scenario
• [MB-1911] - ClassCastException in topic subscription search
• [MB-1912] - NPE thrown when a node with active non-durable subscriptions is killed in a 2 node cluster
• [MB-1914] - Deleted queues are still listed in the sync nodes even though the 'Queue sync [delete]' logged
• [MB-1916] - [Intermittent] Some slots remain in a fail-over scenario
• [MB-1918] - "metadata read from database" printed twice for some messages when failing over

Task

• [MB-1040] - Long running tests and performance test with MySQL
• [MB-1771] - Document Subscription Filtering Feature
• [MB-1852] - Remove MQTT code from WSO2 Message Broker Code
• [MB-1854] - Test and make sure message tracing work for WSO2 MB without restart
• [MB-1883] - Create migration scripts from WSO2MB 3.1.0 to WSO2MB 3.2.0
• [MB-1915] - Remove Hazelcast based slot implementation

Improvement

• [MB-1510] - Queue Subscriptions cannot be identified uniquely
• [MB-1662] - Allow to configure number of reserved db connections for Publisher transactions
• [MB-1733] - Provide subscriber information for Queues/Topics
• [MB-1802] - Improve the UI to provide the number of subscribers connected to the MB
• [MB-1842] - [Intermittent] Task ID is not printed - Error occurred while processing task. Task id
• [MB-1875] - Add internal message ID to the message list UI page