Fault-Tolerant Baseline Application

Plan [ Top ]

1. Tomcat runs on machine 1
2. JNDI runs on machine 1
3. RepMan launches JBoss on machine 2 and machine 3 and registers the two replicas with JNDI
4. Servlet connect to default server (machine 2)
5. Machine 2's server fails
6. Servlet detects exception and informs RepMan of failed machine
7. RepMan removes failed machine from JNDI list
8. Servlet requests next available machine from RepMan
9. RepMan returns machine 3
10. Servlet connects to machine 3 and continues service

1. FD pings all available hosts
   • We must determine how often, and with what timeout
2. FD queries RepMan for available hosts - in case new ones were launched
   • We must determine how often
3. FD detects failure (timeout from a ping)
4. FD requests RepMan to kill failed replica (so the label is removed from JNDI)

1. RepMan launches JBoss on machine 2 and machine 3 and registers the two replicas with JNDI
2. When RepMan is informed of failure from Servlet or Fault Detector, it will launch a new replica and bring the old one back to life
   • We must consider how to cleanly kill JBoss
   • If FD and RepMan's available hosts are not in sync, this might cause duplicate kills and deadlock
   • It takes sometime to start JBoss

1. No state kept in beans, do not need to handle stored state
2. Purchase transaction is the only transaction that can cause difficulty with attempts at duplication
3. Use client username as unique identifier for client transactions of this type

Current Status [ Top ]

Downloads [ Top ]