Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
package com.teambytes.inflatable.raft.protocol
import akka.actor.ActorRef
import scala.collection.immutable
import com.teambytes.inflatable.raft.model.{RaftSnapshot, Entry, ReplicatedLog, Term}
import com.teambytes.inflatable.raft.ClusterConfiguration
private[protocol] trait RaftProtocol extends Serializable {
sealed trait RaftMessage extends Message[Raft]
/**
* Wrap messages you want to send to the underlying replicated state machine
*/
case class ClientMessage[T](client: ActorRef, cmd: T) extends RaftMessage
/**
* Message sent by a [[com.teambytes.inflatable.raft.Candidate]] in order to win an election (and become [[com.teambytes.inflatable.raft.Leader]]).
*
* @see §5.2 Leader Election and §5.4.1 - Election Restriction
*/
case class RequestVote(
term: Term,
candidateId: ActorRef,
lastLogTerm: Term,
lastLogIndex: Int
) extends RaftMessage
case class AppendEntries[T](
term: Term,
prevLogTerm: Term,
prevLogIndex: Int,
entries: immutable.Seq[Entry[T]],
leaderCommitId: Int
) extends RaftMessage {
def isHeartbeat = entries.isEmpty
def isNotHeartbeat = !isHeartbeat
override def toString = s"""AppendEntries(term:$term,prevLog:($prevLogTerm,$prevLogIndex),entries:$entries)"""
}
object AppendEntries {
def apply[T](term: Term, replicatedLog: ReplicatedLog[T], fromIndex: Int, leaderCommitId: Int): AppendEntries[T] = {
val entries = replicatedLog.entriesBatchFrom(fromIndex)
entries.headOption match {
case Some(head) => new AppendEntries[T](term, replicatedLog.termAt(head.prevIndex), head.prevIndex, entries, leaderCommitId)
case _ => new AppendEntries[T](term, Term(1), 1, entries, leaderCommitId)
}
}
}
/**
* Used by Members to notify [[com.teambytes.inflatable.raft.RaftClientActor]] actors that they're sending their requests to a non-leader, and
* that they should update their internal `leader` ActorRef to the given one.
*
* Upon receive of such message, all further communication should be done with the given actorRef.
*
* @see §8 - Client Interaction
*
* @param ref the leader's ActorRef if known by the contacted member (or None if unknown, or the member is just candidating to become the new Leader)
* @param msg if this message is sent in response to a client sending `msg` to a Follower or Candidate,
* they could not take the write, and the client must retry sending this message once it knows the Leader's ActorRef.
* To ease this, the "rejected" message will be carried in this parameter.
*/
case class LeaderIs(ref: Option[ActorRef], msg: Option[Any]) extends Message[Raft]
case object WhoIsTheLeader extends Message[Raft]
/**
* Raft extension: Snapshots
* Used by the Leader to install a snapshot onto an "catching up from very behind" Follower.
* It's the Followers responsibility to load the snapshot data and apply it to it's state machine, by using akka-persistence provided mechanisms.
*
* @see §7 - Log Compaction
*/
case class InstallSnapshot(snapshot: RaftSnapshot) extends Message[Raft]
/**
* Raft extension: Cluster Membership Changes / Joint Consensus
*
* There can take a while to propagate, and will only be applied when the config is passed on to all nodes,
* and the Leader (current, or new) is contained in the new cluster configuration - this may cause a re-election,
* if we're about to remove the current leader from the cluster.
*
* @see §6 - Cluster Membership Changes
*/
case class ChangeConfiguration(newConf: ClusterConfiguration) extends Message[Raft]
/**
* Message issued by freshly restarted actor after has crashed
* Leader reacts with sending [[com.teambytes.inflatable.raft.protocol.RaftProtocol.ChangeConfiguration]]
*
* @see §6 - Cluster Membership Changes
*/
case object RequestConfiguration extends Message[Internal]
}
