When failover.initialReconnectDelay is set, client cannot connect

When option failover.initialReconnectDelay is set to other value than 0, client hangs up during creating a session.

InitialContext ctx = new InitialContext(properties);
Connection connection = ((ConnectionFactory) ctx.lookup("connection")).createConnection();
Session session = connection.createSession(false, Session.CLIENT_ACKNOWLEDGE);

Connection string used:

connectionfactory.connection=failover:(amqp://server:21415?amqp.saslMechanisms=PLAIN)?failover.initialReconnectDelay=1&failover.maxReconnectAttempts=5&failover.reconnectDelay=60000&jms.username=admin&jms.password=admin&jms.forceAsyncSend=false&jms.alwaysSyncSend=true&jms.sendAcksAsync=false

Broker was started day before test and broker still is running. In broker logs there are no signs about connection (trace level) and client is just stuck (i waited for 10 minutes and still no error message).

Broker is qpidd - the Qpid AMQP Message Broker Daemon, versions of qpidd and mrg-m below.

$ qpidd --version
qpidd (qpid-cpp) version 0.22
$ rpm -qa | grep mrg-m
mrg-messaging-release-3.0.0-3.el6.noarch
$ rpm -qa | grep qpid
qpid-cpp-server-0.22-52.el6.x86_64
qpid-cpp-server-linearstore-0.22-52.el6.x86_64
qpid-proton-c-0.7-4.el6.x86_64
qpid-tools-0.22-16.el6.noarch
python-qpid-qmf-0.22-41.el6.x86_64
qpid-cpp-client-devel-0.22-52.el6.x86_64
qpid-cpp-client-devel-docs-0.22-52.el6.noarch
python-qpid-0.22-19.el6.noarch
qpid-cpp-server-devel-0.22-52.el6.x86_64
qpid-qmf-0.22-41.el6.x86_64
qpid-cpp-client-0.22-52.el6.x86_64
qpid-cpp-server-ha-0.22-52.el6.x86_64
qpid-tests-0.22-16.el6.noarch