Description of problem: failover_soak highlights low-level client reconnect failures (exit code of 6). Test failover_soak sometimes (seen on RHEL 5.8 / 6.3 i686) fails with END_OF_TEST ERROR_CLIENT (101/0) It looks that low-level clients in 32 bit environment fails to reconnect... failover_soak logging: soak-0: 2012-10-08 06:03:42 notice cluster(10.16.66.66:29085 READY) joined cluster soakTestCluster_cc3f8332-e3fc-4e6b-b6b9-431a3f9a83aa new broker created: pid == 29085 log-prefix == soak-0 soak-0: 2012-10-08 06:03:42 notice Broker running ... new broker created: pid == 29090 log-prefix == soak-1 Broker List ------------ size: 2 pid: 29085 port: 41347 pid: 29090 port: 49818 end Broker List ------------ 2012-10-08_06:03:43 Declaring queues... ... 2012-10-08_06:08:05 Killing oldest broker... killFrontBroker: waiting for newbie sync... killFrontBroker: newbie synced. killFrontBroker pid: 29085 on port 41347 using signal 9 ... new broker created: pid == 29122 log-prefix == soak-2 Broker List ------------ size: 2 pid: 29090 port: 49818 pid: 29122 port: 49797 end Broker List ------------ 2012-10-08_06:09:08 1 brokers have been killed. ... new broker created: pid == 29137 log-prefix == soak-3 Broker List ------------ size: 2 pid: 29122 port: 49797 pid: 29137 port: 43999 end Broker List ------------ 2012-10-08_06:10:02 2 brokers have been killed. ... new broker created: pid == 29155 log-prefix == soak-4 Broker List ------------ size: 2 pid: 29137 port: 43999 pid: 29155 port: 45541 end Broker List ------------ 2012-10-08_06:15:12 3 brokers have been killed. ... new broker created: pid == 29170 log-prefix == soak-5 Broker List ------------ size: 2 pid: 29155 port: 45541 pid: 29170 port: 38701 end Broker List ------------ 2012-10-08_06:16:09 4 brokers have been killed. ... new broker created: pid == 29185 log-prefix == soak-6 Broker List ------------ size: 2 pid: 29170 port: 38701 pid: 29185 port: 46375 end Broker List ------------ 2012-10-08_06:17:14 5 brokers have been killed. ... new broker created: pid == 29200 log-prefix == soak-7 Broker List ------------ size: 2 pid: 29185 port: 46375 pid: 29200 port: 51957 end Broker List ------------ 2012-10-08_06:20:31 6 brokers have been killed. ... new broker created: pid == 29214 log-prefix == soak-8 Broker List ------------ size: 2 pid: 29200 port: 51957 pid: 29214 port: 59325 end Broker List ------------ 2012-10-08_06:20:59 7 brokers have been killed. Sender logging: SENDER started (PID:29105) 2012-10-08 06:03:55 info ... ... SENDER (host: 127.0.0.1 port: 41347 ) PID: 29105 Failed: Cannot establish a connection 2012-10-08 06:20:58 info ... Broker does not show any suspicious messages. Version-Release number of selected component (if applicable): qpid-cpp-client-0.14-22.el6_3.i686 qpid-cpp-client-devel-0.14-22.el6_3.i686 qpid-cpp-client-devel-docs-0.14-22.el6_3.noarch qpid-cpp-client-rdma-0.14-22.el6_3.i686 qpid-cpp-client-ssl-0.14-22.el6_3.i686 qpid-cpp-debuginfo-0.14-22.el6_3.i686 qpid-cpp-server-0.14-22.el6_3.i686 qpid-cpp-server-cluster-0.14-22.el6_3.i686 qpid-cpp-server-devel-0.14-22.el6_3.i686 qpid-cpp-server-rdma-0.14-22.el6_3.i686 qpid-cpp-server-ssl-0.14-22.el6_3.i686 qpid-cpp-server-store-0.14-22.el6_3.i686 qpid-cpp-server-xml-0.14-22.el6_3.i686 qpid-java-client-0.18-2.el6.noarch qpid-java-common-0.18-2.el6.noarch qpid-java-example-0.18-2.el6.noarch qpid-qmf-0.14-14.el6_3.i686 qpid-qmf-debuginfo-0.14-14.el6_3.i686 qpid-qmf-devel-0.14-14.el6_3.i686 qpid-tests-0.14-1.el6_2.noarch qpid-tools-0.14-6.el6_3.noarch python-qpid-0.14-11.el6_3.noarch python-qpid-qmf-0.14-14.el6_3.i686 How reproducible: 5% Steps to Reproduce: 1. run failover_soak in loop as qpid_ptest_cluster_failover_soak does 2. watch for exit codes == 6 Actual results: failover_soak fails. Expected results: failover_soak should pass. Additional info:
Created attachment 623470 [details] failover_soak and broker logs
Frantisek, retest on 0.18.
This is medium (or perhaps low) priority because it affects only the old client api, not the new messaging api. Moving out to the next release.
The current runs on 0.18 based qpid show that issue is not occurring anymore, dependency to bug 875660. Will retest workarounding bug 875660 or eventually after bug 875660 is fixed.
(In reply to comment #6) > The current runs on 0.18 based qpid show that issue is not occurring > anymore, dependency to bug 875660. Will retest workarounding bug 875660 or > eventually after bug 875660 is fixed. Okay, moving this to modified for now.
The bug 875660 has been verified, but we still see the initial issue occurring sporadically on the latest 0.18 builds. -> ASSIGNED
Since this is a bug against the deprecated client api, -> wontfix