[ https://issues.apache.org/jira/browse/DISPATCH-380?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ted Ross updated DISPATCH-380: ------------------------------ Fix Version/s: (was: 0.6.0) > Router stops receiving messages from multiple senders publishing to multiple > queues in parallel > ----------------------------------------------------------------------------------------------- > > Key: DISPATCH-380 > URL: https://issues.apache.org/jira/browse/DISPATCH-380 > Project: Qpid Dispatch > Issue Type: Bug > Components: Routing Engine > Environment: Debian 8.3, Apache Qpid Proton 0.13.0-RC for drivers and > dependencies, Hardware: 2 CPUs, 15 GB RAM, 60 GB HDD each on 3 separate > machines > Reporter: Vishal Sharda > Priority: Critical > Attachments: AWS_hung_at_round_figures.png, > AWS_hung_for_4_seconds.png, AWS_uneven_start.png, Senders_1.png, > Senders_2.png, Senders_3_10_Queues.png, Senders_4_10_queues.png, > Single_router_testing_results.pdf, qdstat_wrong_output.png > > > I am running a Java Client against a cluster of 3 interior routers connected > to each other. 2-way SSL is enabled for all the connections. > There were 20 simultaneous queues with 20 senders on each queue and each > sender publishing 1000 messages. All the senders were connected to Router 1. > 20 receivers were connected to Router 2 with 1 receiver receiving from each > queue. > In the first run, router stopped receiving incoming messages after delivering > 386,339 out of 400K "Hello World!" messages. > In the second run, 388,781 messages out of 400K were delivered. > I reduced the number of queues to 10 (halving total number of messages to > 200K) and the issue occurred again. > I ran the Java client on an 8 CPU machine again with 10 queues and the issue > occurred again after delivering just 54K out of 200K messages. > All the senders were hung (still connected) with no messages flowing at all. > Connection information from qdstat: > When the messages are flowing properly and I run "qdstat -c", I see all the > senders as secure and authenticated. > After they hang and I run "qdstat -c", it erroneously shows all the clients > as insecure and unauthenticated. > Shortly after the clients hang, all the queues are deleted from the router > network but connections are still shown until I terminate the clients. > I saw this erroneous situation before also when "qdstat -c" showed some > senders as secure and authentic but some as insecure/unauthentic. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org For additional commands, e-mail: dev-h...@qpid.apache.org