Persistent Messaging – Store on File system or RDBMS – Survive restart, and process failure • Master / Slave – Lock manager – Shared Storage - SAN/GFS2 or NFS v4 or RDBMS – Replicated - Block or RDBMS replication • Managed – Red Hat Cluster Suite – Shared or Replicated storage
– tcp://<host>:<port>?<connection options> – Simple, but no reconnection on failure • Failover Transport – failover:(tcp://<host>:<port>,tcp://<host>:<port>,...) – Reconnect on failure • Fabric Discovery – discovery:(fabric://<group name>) – Reconnect of failure – Coordinates with Fabric Registry
• Smarter failure management – N+1 failover architecture – Restart on node first, then failover to back node – More sophisticated detection than just process failure • RHEL only • Complements JBoss A-MQ HA capabilities • JUDCon Boston 2013 – Resilient Messaging – http://www.jboss.org/dms/judcon/2013unitedstates/presentati ons/judcon2013_day3track3session5.pdf – http://bit.ly/18kcidy
a highly scalable and available open source messaging product • Many HA options to meet your needs • Links – http://www.jboss.org/products/amq – https://github.com/FuseByExample/external-mq-fabric-client/ – JUDCon Boston 2013 Presentation - http://bit.ly/18kcidy