rsyslog messages are lost when failing over rsyslog server

Issue

• rsyslog messages are lost when failing over rsyslog server
• I have built a 2 node HA cluster using pacemaker, cman and corosync. This supports an instance of HAProxy in Active/Active mode supporting rsyslog. The supporting VIP is managed by iptables and clusterIP. When running rsyslog over RELP and one node is deactivated there is no loss of message traffic, which is expected. When running rsyslog over TCP and one node is deactivated the client takes approx 3 secs to fail over.
• The time between rsyslog stopping on one server and the failover server's ARP for the Virtual IP results in loss of rsyslog messages