Effective October 27, 2012, online and email support for FuseSource products will move to Red Hat support channels. For more information, please see the JIRA Migration to Red Hat FAQ.
As of October 27th, please open all new issues in the Red Hat Customer Portal .
Issue Details (XML | Word | Printable)

Key: ESB-1574
Type: Bug Bug
Status: Open Open
Priority: Major Major
Assignee: Jonathan Anstey
Reporter: Jakub Korab
Votes: 0
Watchers: 1
Operations

If you were logged in you would be able to see more operations.
FUSE ESB

Request-response messaging over temp queues locks up on load from multiple producer bundle instances

Created: 20/Oct/11 02:49 PM   Updated: 25/Oct/11 04:14 PM
Component/s: None
Affects Version/s: 4.4.1-fuse-00-08
Fix Version/s: None

File Attachments: 1. Zip Archive ping-pong.zip (5 kB)

Environment: Windows 7 server, Macbook Air
Issue Links:
Blockers
 

External Issue URL: http://fusesource.com/issues/browse/ESB-1573


 Description  « Hide
Two bundles are used to test this (attached in ESB-1573):
  • a producer (pinger) that makes a request via a Camel inOut("activemq:...") and waits for a response
  • a consumer (ponger) that pulls from that queue and responds on a temporary queue

Two Smx instances are set up with the pinger bundle, one is set up with the ponger, and tied together through a single Amq instance. A load test is set up using JMeter to hit the ping web service with 100 threads repeatedly, and run against both Smx instances simultaneously. After a handful of requests, the producer bundles start throwing exceptions about exceeded timeouts on the inOut() segment.

No exceptions are seen from the ponger bundle.



 All   Comments   Change History      Sort Order: Ascending order - Click to sort in descending order
Jakub Korab made changes - 20/Oct/11 02:52 PM
Field Original Value New Value
Description Two bundles are used to test this (attached in MB-934):
- a producer (pinger) that makes a request via a Camel inOut("activemq:...") and waits for a response
- a consumer (ponger) that pulls from that queue and responds on a temporary queue

Two Smx instances are set up with the pinger bundle, one is set up with the ponger, and tied together through a single Amq instance. A load test is set up using JMeter to hit the ping web service with 100 threads repeatedly, and run against both Smx instances simultaneously. After a handful of requests, the producer bundles start throwing exceptions about exceeded timeouts on the inOut() segment.

No exceptions are seen from the ponger bundle.
Two bundles are used to test this (attached in ESB-1573):
- a producer (pinger) that makes a request via a Camel inOut("activemq:...") and waits for a response
- a consumer (ponger) that pulls from that queue and responds on a temporary queue

Two Smx instances are set up with the pinger bundle, one is set up with the ponger, and tied together through a single Amq instance. A load test is set up using JMeter to hit the ping web service with 100 threads repeatedly, and run against both Smx instances simultaneously. After a handful of requests, the producer bundles start throwing exceptions about exceeded timeouts on the inOut() segment.

No exceptions are seen from the ponger bundle.
Jonathan Anstey made changes - 20/Oct/11 04:41 PM
Comment [ This is likely due to MB-934, which isn't fixed in any release yet. There are instruction in that issue to try a SNAPSHOT version of AMQ that has the fix. ]
Jonathan Anstey made changes - 25/Oct/11 04:06 PM
Assignee Jonathan Anstey [ janstey ]
Jonathan Anstey made changes - 25/Oct/11 04:07 PM
Link This issue is blocked by ESB-1573 [ ESB-1573 ]
Claus Ibsen made changes - 25/Oct/11 04:14 PM
Attachment ping-pong.zip [ 20828 ]