hbase replication原理分析

本文只是从总体流程来分析replication过程，很多细节没有提及，下一篇文章准备多分析分析细节。

 

replicationSource启动过程

org.apache.hadoop.hbase.regionserver.HRegionServer#startServiceThreads ->

org.apache.hadoop.hbase.replication.regionserver.Replication#startReplicationService ->

 //初始化replicationManager

org.apache.hadoop.hbase.replication.regionserver.ReplicationSourceManager#init ->

//在init阶段for循环把所有的replicationPeers添加到source里，即每个replicationPeer对应一个source，也就是可以添加多个slave cluster，replicationPeers从zookeeper /hbase/replication/peers目录取

org.apache.hadoop.hbase.replication.regionserver.ReplicationSourceManager#addSource ->

//在addSource阶段生成ReplicationSource并启动ReplicationSource，ReplicationSource本身是一个线程

org.apache.hadoop.hbase.replication.regionserver.ReplicationSource#startup

//ReplicationSource线程启动，进入while循环工作

 

 

replicationSource大致工作流程

1. while(isAlive())进行主体循环
2. 从WAL文件获取List<WAL.Entry>
3. 通过调用shipEdits方法发送数据
4. 调用replicationEndpoint replicate方法发送数据
5. 最终调用admin.replicateWALEntry通过rpc发送数据

 

regionserver如何从slave cluster中选取regionserver当做复制节点

1. replication过程需要连接peer（slave cluster），首先要获取这个peer所有活着的regionservers
2. 拿到所有regionservers信息之后，开始选择哪些regionservers作为replication的对象
3. 选哪些regionservers当做sink由peer活着的regionserver个数*ratio（默认值0.1）决定，regionservers先shuffle打乱顺序后再截取
4. 如果选择的sink（regionserver）个数为0，一直等待peer上线，也就是slave cluster没有启动的情况
5. 下面源码可以解释如何选择regionserver当做sink

  private void connectToPeers() {
    getRegionServers();

    int sleepMultiplier = 1;

    // Connect to peer cluster first, unless we have to stop
    while (this.isRunning() && replicationSinkMgr.getSinks().size() == 0) {
      replicationSinkMgr.chooseSinks();
      if (this.isRunning() && replicationSinkMgr.getSinks().size() == 0) {
        if (sleepForRetries("Waiting for peers", sleepMultiplier)) {
          sleepMultiplier++;     //倍数最多为默认配置的300倍，也就是每次sleep最长间隔是300秒
        }
      }
    }
  }

  void chooseSinks() {
    List<ServerName> slaveAddresses = endpoint.getRegionServers();
    Collections.shuffle(slaveAddresses, random);
    int numSinks = (int) Math.ceil(slaveAddresses.size() * ratio);
    sinks = slaveAddresses.subList(0, numSinks);
    lastUpdateToPeers = System.currentTimeMillis();
    badReportCounts.clear();
  }

  /**
   * Do the sleeping logic
   * @param msg Why we sleep
   * @param sleepMultiplier by how many times the default sleeping time is augmented
   * @return True if <code>sleepMultiplier</code> is &lt; <code>maxRetriesMultiplier</code>
   */
  protected boolean sleepForRetries(String msg, int sleepMultiplier) {
    try {
      if (LOG.isTraceEnabled()) {
        LOG.trace(msg + ", sleeping " + sleepForRetries + " times " + sleepMultiplier);
      }
      Thread.sleep(this.sleepForRetries * sleepMultiplier);
    } catch (InterruptedException e) {
      LOG.debug("Interrupted while sleeping between retries");
    }
    return sleepMultiplier < maxRetriesMultiplier;
  }

this.maxRetriesMultiplier = this.conf.getInt("replication.source.maxretriesmultiplier", 300);
this.ratio = conf.getFloat("replication.source.ratio", DEFAULT_REPLICATION_SOURCE_RATIO);

    总结

1. 每个slave cluster对应一个replicationSource线程，各个slave复制互不干扰
2. 每个replicationSource是单线程进行传输数据，改成多线程并发传可能更好
3. 数据是通过rpc发送过去，调用slave cluster regionserver RSRpcServices的replicateWALEntry方法