-
Bug
-
Resolution: Done
-
Blocker
-
7.0.0.ER2 (Beta)
-
None
There is lost large message in cluster with network failures. This can have negative impact for customers.
Test scenario:
1. Start 2 servers in Artemis cluster with deployed "queue/InQueue" (reconnect-attempts for cluster connection is set to -1)
2. Start producer which sends large messages (1 MB) to InQueue to 1st server
3. Start consumer on 2nd server which reads messages from InQueue
4. During steps 2. and 3. disconnect network between servers, wait 2 minutes and reconnect
5. Stop producer and wait for consumer to receive all messages
6. Verify that number of send and received messages is equal
There is only one record in race log about lost message on receiving server:
node-2-log/server-trace.log:09:58:21,123 TRACE [org.apache.activemq.artemis.core.server] (default I/O-4) sendLarge::LargeServerMessage[messageID=1279,priority=4,expiration=[null], durable=true, address=jms.queue.InQueue,properties=TypedProperties[__AMQ_CID=f77d9fb2-98d1-11e5-b811-01a8b3bcda7a,count=235,color=RED,_AMQ_BRIDGE_DUP=[E0A6 E941 98D1 11E5 B192 6118 E771 B016 0000 0000 0000 05F8),counter=236,_AMQ_DUPL_ID=45afdb1a-8aa8-4ff6-ae7d-44ed84e4a98b1449046516079,_AMQ_LARGE_SIZE=1048576,_AMQ_ROUTE_TO=[0000 0000 0000 000C),bytesAsLongs(12]]]@1469761220
- is blocked by
-
JBEAP-2499 Upgrade Artemis to 1.1.0.wildfly-011
- Closed
- relates to
-
JBEAP-2315 Lost messages in XA transaction between JMS servers in cluster and MDBs
- Closed
-
JBEAP-2544 Duplicate delivery over Bridges under OME scenarios, paging and other failures
- Closed