QPID-5720: HA exception raised by the store for durable transactions

Running qpid-txtest against a HA cluster was raising this exception:
async_dequeue() failed: jexception 0x0b02 wmgr::dequeue_check() threw JERR_MAP_LOCKED: Record ID locked by a pending transaction.

This is actually a test bug. In a cluster a transaction commit takes longer to
complete because of co-ordinating with backups, the test was not waiting for the
completion of commit command before proceeding. Adding a sync() in the test
solves the problem.  Note this test uses the old obsolete C++ API. Applications
written using the supported messaging API are not affected by this. This can be
verified using the qpid-send and qpid-receive clients with the --tx option
(there is already an automated test for this in ha_tests.py.)

git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1589520 13f79535-47bb-0310-9956-ffa450edef68

1 files changed, 1 insertions, 0 deletions

diff --git a/qpid/cpp/src/tests/qpid-txtest.cpp b/qpid/cpp/src/tests/qpid-txtest.cpp
index 6e7d46802c..59ab905af7 100644
--- a/qpid/cpp/src/tests/qpid-txtest.cpp
+++ b/qpid/cpp/src/tests/qpid-txtest.cpp
@@ -179,6 +179,7 @@ struct Transfer : public Client, public Runnable
                 } else {
                     session.txCommit();
                 }
+                session.sync();
             }
         } catch(const std::exception& e) {
             std::cout << "Transfer interrupted: " << e.what() << std::endl;