Package org.opendaylight.controller.cluster.raft.behaviors

Class Leader

java.lang.Object
org.opendaylight.controller.cluster.raft.behaviors.AbstractRaftActorBehavior
org.opendaylight.controller.cluster.raft.behaviors.AbstractLeader
org.opendaylight.controller.cluster.raft.behaviors.Leader
All Implemented Interfaces:
AutoCloseable, RaftActorBehavior
public class Leader extends AbstractLeader
The behavior of a RaftActor when it is in the Leader state.

Leaders:

  • Upon election: send initial empty AppendEntries RPCs (heartbeat) to each server; repeat during idle periods to prevent election timeouts (§5.2)
  • If command received from client: append entry to local log, respond after entry applied to state machine (§5.3)
  • If last log index ≥ nextIndex for a follower: send AppendEntries RPC with log entries starting at nextIndex
  • If successful: update nextIndex and matchIndex for follower (§5.3)
  • If AppendEntries fails because of log inconsistency: decrement nextIndex and retry (§5.3)
  • If there exists an N such that N > commitIndex, a majority of matchIndex[i] ≥ N, and log[N].term == currentTerm: set commitIndex = N (§5.3, §5.4).

  • Constructor Details

  • Method Details

    • handleMessage

      public RaftActorBehavior handleMessage(ActorRef sender, Object originalMessage)
      Description copied from interface: RaftActorBehavior
      Handle a message. If the processing of the message warrants a state change then a new behavior should be returned otherwise this method should return the current behavior.
      Specified by:
      handleMessage in interface RaftActorBehavior
      Overrides:
      handleMessage in class AbstractLeader
      Parameters:
      sender - The sender of the message
      originalMessage - A message that needs to be processed
      Returns:
      The new behavior or current behavior, or null if the message was not handled.

    • beforeSendHeartbeat

      protected void beforeSendHeartbeat()
      Overrides:
      beforeSendHeartbeat in class AbstractLeader

    • handleAppendEntriesReply

      protected RaftActorBehavior handleAppendEntriesReply(ActorRef sender, AppendEntriesReply appendEntriesReply)
      Description copied from class: AbstractRaftActorBehavior
      Derived classes should not directly handle AppendEntriesReply messages it should let the base class handle it first. Once the base class handles the AppendEntriesReply message and does the common actions that are applicable in all RaftState's it will delegate the handling of the AppendEntriesReply message to the derived class to do more state specific handling by calling this method
      Overrides:
      handleAppendEntriesReply in class AbstractLeader
      Parameters:
      sender - The actor that sent this message
      appendEntriesReply - The AppendEntriesReply message
      Returns:
      a new behavior if it was changed or the current behavior

    • transferLeadership

      public void transferLeadership(@NonNull RaftActorLeadershipTransferCohort leadershipTransferCohort)
      Attempts to transfer leadership to a follower as per the raft paper (§3.10) as follows:
      • Start a timer (Stopwatch).
      • Send an initial AppendEntries heartbeat to all followers.
      • On AppendEntriesReply, check if the follower's new match Index matches the leader's last index
      • If it matches,
      • Otherwise if the election time out period elapses, notify RaftActorLeadershipTransferCohort.abortTransfer().
      Parameters:
      leadershipTransferCohort - the cohort participating in the leadership transfer

    • close

      public void close()
      Specified by:
      close in interface AutoCloseable
      Specified by:
      close in interface RaftActorBehavior
      Overrides:
      close in class AbstractLeader