// sendAppend sends RRPC, with entries to the given peer.
func (r *raft) sendAppend(to uint64) {
pr := r.prs[to]
if pr.isPaused() {
return
}
m := pb.Message{}
m.To = to
term, errt := r.raftLog.term(pr.Next - 1)
ents, erre := r.raftLog.entries(pr.Next, r.maxMsgSize)
if errt != nil || erre != nil { // send snapshot if we failed to get term or entries
m.Type = pb.MsgSnap
snapshot, err := r.raftLog.snapshot()
if err != nil {
if err == ErrSnapshotTemporarilyUnavailable {
r.logger.Debugf("%x failed to send snapshot to %x because snapshot is temporarily unavailable", r.id, to)
return
}
panic(err) // TODO(bdarnell)
}
if IsEmptySnap(snapshot) {
panic("need non-empty snapshot")
}
m.Snapshot = snapshot
sindex, sterm := snapshot.Metadata.Index, snapshot.Metadata.Term
r.logger.Debugf("%x [firstindex: %d, commit: %d] sent snapshot[index: %d, term: %d] to %x [%s]",
r.id, r.raftLog.firstIndex(), r.Commit, sindex, sterm, to, pr)
pr.becomeSnapshot(sindex)
r.logger.Debugf("%x paused sending replication messages to %x [%s]", r.id, to, pr)
} else {
m.Type = pb.MsgApp
m.Index = pr.Next - 1
m.LogTerm = term
m.Entries = ents
m.Commit = r.raftLog.committed
if n := len(m.Entries); n != 0 {
switch pr.State {
// optimistically increase the next when in ProgressStateReplicate
case ProgressStateReplicate:
last := m.Entries[n-1].Index
pr.optimisticUpdate(last)
pr.ins.add(last)
case ProgressStateProbe:
pr.pause()
default:
r.logger.Panicf("%x is sending append in unhandled state %s", r.id, pr.State)
}
}
}
r.send(m)
}
func (dec *msgAppV2Decoder) decode() (raftpb.Message, error) {
var (
m raftpb.Message
typ uint8
)
if _, err := io.ReadFull(dec.r, dec.uint8buf); err != nil {
return m, err
}
typ = uint8(dec.uint8buf[0])
switch typ {
case msgTypeLinkHeartbeat:
return linkHeartbeatMessage, nil
case msgTypeAppEntries:
m = raftpb.Message{
Type: raftpb.MsgApp,
From: uint64(dec.remote),
To: uint64(dec.local),
Term: dec.term,
LogTerm: dec.term,
Index: dec.index,
}
// decode entries
if _, err := io.ReadFull(dec.r, dec.uint64buf); err != nil {
return m, err
}
l := binary.BigEndian.Uint64(dec.uint64buf)
m.Entries = make([]raftpb.Entry, int(l))
for i := 0; i < int(l); i++ {
if _, err := io.ReadFull(dec.r, dec.uint64buf); err != nil {
return m, err
}
size := binary.BigEndian.Uint64(dec.uint64buf)
var buf []byte
if size < msgAppV2BufSize {
buf = dec.buf[:size]
if _, err := io.ReadFull(dec.r, buf); err != nil {
return m, err
}
} else {
buf = make([]byte, int(size))
if _, err := io.ReadFull(dec.r, buf); err != nil {
return m, err
}
}
dec.index++
// 1 alloc
pbutil.MustUnmarshal(&m.Entries[i], buf)
}
// decode commit index
if _, err := io.ReadFull(dec.r, dec.uint64buf); err != nil {
return m, err
}
m.Commit = binary.BigEndian.Uint64(dec.uint64buf)
case msgTypeApp:
var size uint64
if err := binary.Read(dec.r, binary.BigEndian, &size); err != nil {
return m, err
}
buf := make([]byte, int(size))
if _, err := io.ReadFull(dec.r, buf); err != nil {
return m, err
}
pbutil.MustUnmarshal(&m, buf)
dec.term = m.Term
dec.index = m.Index
if l := len(m.Entries); l > 0 {
dec.index = m.Entries[l-1].Index
}
default:
return m, fmt.Errorf("failed to parse type %d in msgappv2 stream", typ)
}
return m, nil
}