关于RAFT后续的一些思考(二)

这篇是关于优化的snapshot。从连接中2.6小节快照管理Raft 算法原理及其在 CMQ 中的应用（上）中描述的那样，在floyd中实现也比较简单：

313     } else if (res.append_entries_res().success() == true) {
314       if (num_entries > 0) {
315         match_index_ = prev_log_index + num_entries;
316         // only log entries from the leader's current term are committed
317         // by counting replicas
318         if (append_entries->entries(num_entries - 1).term() == context_->current_term) {
319           AdvanceLeaderCommitIndex();
320           apply_->ScheduleApply();
321         }
322         next_index_ = prev_log_index + num_entries + 1;
323       }

194 // only leader will call AdvanceCommitIndex
195 // follower only need set commit as leader's
196 void Peer::AdvanceLeaderCommitIndex() {
197   Entry entry;
198   uint64_t new_commit_index = QuorumMatchIndex();
199   if (context_->commit_index < new_commit_index) {
200     context_->commit_index = new_commit_index;
201     raft_meta_->SetCommitIndex(context_->commit_index);
202   }
203   return;
204 }

 62 void FloydApply::ApplyStateMachine() {
 63   uint64_t last_applied = context_->last_applied;
 64   // Apply as more entry as possible
 65   uint64_t commit_index;
 66   commit_index = context_->commit_index;
 67 
 70   Entry log_entry;
 71   if (last_applied >= commit_index) {
 72     return;
 73   }
 74   // TODO: use batch commit to optimization
 75   while (last_applied < commit_index) {
 76     last_applied++;
 77     raft_log_->GetEntry(last_applied, &log_entry);
 78     // TODO: we need change the s type
 79     // since the Apply may not operate rocksdb
 80     rocksdb::Status s = Apply(log_entry);
 81     if (!s.ok()) {
 84       usleep(1000000);
 85       ScheduleApply();  // try once more
 86       return;
 87     }
 88   }
 89   context_->apply_mu.Lock();
 90   context_->last_applied = last_applied;
 91   raft_meta_->SetLastApplied(last_applied); ///持久化applyindex
 92   context_->apply_mu.Unlock();
 93   context_->apply_cond.SignalAll();
 94 }

参考资料
Raft 算法原理及其在 CMQ 中的应用（下