Failure Detection System To-Do

[jyotsana-gupta]

In replica1.servers.RMFailDetectUDPThread, code for sending and receiving heartbeats from other RM's. Use ports different from the ones already used. Same for replica2 and replica3.

In frontend.services.RecordManagerFEImpl, in crashLeadServer(), code for simulating lead server crash. This should stop RMFailDetectUDPThread and ReplicaMgrUDPServer for the leader, elect a new leader, send all required notifications and try to restart the stopped threads.

In frontend.services.RecordManagerFEImpl, in crashSecondaryServer(), code for simulating any secondary server crash. This should stop RMFailDetectUDPThread and ReplicaMgrUDPServer for the secondary RM and try to restart the stopped threads.

When a new leader is elected, following methods need to be called:
~ frontend.services.RecordManagerFEImpl.setLeadServerDetails(String leadServerDetails): leadServerDetails will hold "leaderdetails__" of the new leader

~ replica1.services.impl.ReplicaMgrService.setIsLeader(String rmStatusStr): rmStatusStr will hold "leaderstatus_true" for the new leader and "leaderstatus_false" for all secondary RM's

~ replica1.services.impl.FIFOBroadcastSys.setSecServerDetails(List<String[]> secServerDetails): secServerDetails will hold a list of arrays (each array will have 2 elements: hostname and port number). These arrays will give details of the secondary RM's when a new leader is elected.

In failuredetectionsys package, code for any other failure detection activities that are not included in the above classes, such as electing new leader, sending leader details, restarting failed RMFailDetectUDPThread and ReplicaMgrUDPServer.