-
Bug
-
Resolution: Done
-
Blocker
-
7.1.0.DR14, 7.1.0.DR15
-
None
Customer impact: If server in cluster crashes then some messages might get lost.
Test scenario:
- Start two EAP 7.1 servers in (Artemis) cluster
- Send 10 000 messages to queue InQueue to node-1. Wait until all messages are sent.
- Set _AMQ_DUP_ID to generated UUID for each message
- Deploy MDB to node-2 which consumes messages from InQueue, creates new message and sends it to OutQueue (in XA transaction)
- Set inMessageId property of newly created message to message ID of message received from InQueue
- Also _AMQ_DUP_ID of newly created message is set to the same value as for message received from InQueue so they're the same
- Repeat following sequence 5 times
- Kill node-2 and start it after 20 seconds
- Kill node-1 and start it after 20 seconds
- Receive all messages from OutQueue from node -2 and compare message ID of sent messages with inMessageId of received messages to find lost and duplicated messages.
Pass Criteria: Number of messages sent to InQueue will be the same as number of messages received from OutQueue. There are no lost or duplicated messages.
Actual Result:
Sometimes happens that some messages are lost.
Expected Result:
No messages will be lost.
- is blocked by
-
JBEAP-10454 Upgrade Artemis 1.5.4.jbossorg-003
- Closed
- is cloned by
-
ENTMQBR-607 Lost messages in cluster with server crashes
- Closed
- is duplicated by
-
JBEAP-10062 Lost paged messages if journal is located on NFS
- Closed
-
JBEAP-10393 Lost message in BytemanLodh2TestCase.testLodh3KillOnTransactionPrepare
- Closed
-
JBEAP-10475 IndexOutOfBoundsException during reading message if compacting is enabled
- Closed
- is related to
-
JBEAP-10062 Lost paged messages if journal is located on NFS
- Closed
-
JBEAP-10393 Lost message in BytemanLodh2TestCase.testLodh3KillOnTransactionPrepare
- Closed
-
JBEAP-10475 IndexOutOfBoundsException during reading message if compacting is enabled
- Closed
- relates to
-
ARTEMIS-1114 Loading...