从Stage角度看cassandra write

时间:2022-05-27 04:49:08

声明

文章发布于CSDN

cassandra concurrent 具体实现

cassandra并发技术文中介绍了java的concurrent实现,这里介绍cassandra如何基于java实现cassandra并发包。

Figure1——cassandra并发实现
从Stage角度看cassandra write

  1. cassandra各个Stage是通过StageManger来进行管理的,StageManager 有个内部类ExecuteOnlyExecutor。

  2. ExecuteOnlyExecutor继承了ThreadPoolExecutor,实现了cassandra的LocalAwareExecutorSerivce接口

  3. LocalAwareExecutorService继承了Java的ExecutorService,构建了基本的任务模型。添加了两个自己的方法.
    execute方法用于trace跟踪。

    public void execute(Runnable command, ExecutorLocals locals);
    public void maybeExecuteImmediately(Runnable command);

    对于Executor中的默认execute方法,和LocalAwareExecutorSerive中的execute方法都是new 一个task,然后将task添加到queue中。而maybeExecuteImmedicatly方法则是判断下是否有正在执行的task或者work,如果没有则直接执行,而不添加到队列中。

    public void maybeExecuteImmediately(Runnable command)
    {
    //comment1
    FutureTask<?> ft = newTaskFor(command, null);
    if (!takeWorkPermit(false))
    {
    addTask(ft);
    }
    else
    {
    try
    {
    ft.run();
    }
    finally
    {
    returnWorkPermit();
    maybeSchedule();
    }
    }
    }
  4. AbstractLocalAwareExecutorService实现LocalAwareExecutorSerive接口,提供了executor的实现以及ExecutorServie接口中的关于生命周期管理的方法实现,如submit,shoudown等方法。添加了addTask,和任务完成的方法onCompletion。

  5. SEPExecutor实现了LocalAwareExecutorService类,提供了addTask,onCompletion,maybeExecuteImmediately等方法的实现。同时负责队列的管理

  6. SharedExecutorPool,线程池管理,用来管理Executor

cassandra write

cassandra写操作涉及到MutationStage,FlushWriter,MemtablePostFlusher,ReplicateOnWriteStage

MutationStage

  • Single
    Coordinator:将request同时发给所有replicate节点
    Replicate:
    1.写数据到commitlog
    2. 写数据到MemTable
    3. 如果写操作是个delete操作,在commitlog和MemTable中添加墓碑tombstone
    4. 如果使用了row caching,需要失效这行的缓存
    5. 发送应答request到coordinator
    Mutation Stage
    Thread1
    从Stage角度看cassandra write

Thread2:写commit log 和memtable

从Stage角度看cassandra write

  • View/Batch

    Figure2 cassandra mutation change(coordinator)
    从Stage角度看cassandra write
    从Stage角度看cassandra write
    cassandra mutation时序图如上图所示。前面几个都是线程调用和request的”翻译”重点是最后一个类的执行StorageProxy.在#comment1处,cassandra对batch change 和涉及到view更新 与单条的insert操作进行了区分。

    View是和Table绑定在一起的,所以要确保两者是一起更新的。cassandra通过batch log 来实现。无论write consistency level 是多少,batch log 要确保change写入到了quorum份replicate.

    Coordinator:创建batch log,确保quorum份replicate node写入change,客户端的响应仍然是按照write consistency level;将request同时发给所有replicate 节点。
    Replicate:完整调用栈见Figure3

    1. 获得partition 的锁,确保batch/view 的write request是串行化(BatchManger.store#comment1 处)
    2. 如果是视图,则需要读取partition 数据,生成物化视图的增量变化
    3. 写 commit log
    4. 生成batch log
    5. 存储batch log
    6. 发送batch的second write/物化视图的更新到相应的replicate node。因为batch/视图更新 等多条记录可能不在同一个replicate上
    7. 写MemTable
    8. 其他record的replicate会写更新,然后发送response到first record replicate node 上
      //BatchManager.store()

      public static void store(Batch batch, boolean durableWrites) {
      RowUpdateBuilder builder = new RowUpdateBuilder(SystemKeyspace.Batches, batch.creationTime, batch.id)
      .clustering().add("version", MessagingService.current_version);

      for (ByteBuffer mutation : batch.encodedMutations)
      builder.addListEntry("mutations", mutation);

      for (Mutation mutation : batch.decodedMutations) {
      try (DataOutputBuffer buffer = new DataOutputBuffer()) {
      //comment1 串行化多个mutation改变
      Mutation.serializer.serialize(mutation, buffer, MessagingService.current_version);
      builder.addListEntry("mutations", buffer.buffer());
      } catch (IOException e) {
      // shouldn't happen
      throw new AssertionError(e);
      }
      }

      builder.build().apply(durableWrites);
      }

      Figure3 cassandra mutation change(replicate)
      从Stage角度看cassandra write

FlushWriter&&MemtablePostFlusher
FlushWriter Stage 就是将数据从MemTable flush到SSTable中。有三种事件会导致发生
1. memtable 超过了设定的大小
2. nodetool flush
3. commit log 超过了设定的大小
ColumnFamilyStore.switchMemtable方法将Memtable flush到SSTable中.

public ListenableFuture<CommitLogPosition> switchMemtable()
{
synchronized (data)
{
logFlush();
Flush flush = new Flush(false);
flushExecutor.execute(flush);
ListenableFutureTask<CommitLogPosition> task = ListenableFutureTask.create(flush.postFlush);
postFlushExecutor.submit(task);
return task;
}
}

flushExecutor,postFlushExecutor都JMX相关的ThreadPool,因为需要将相关的metrics通过JMX暴露出去

flushExecutor = new JMXEnabledThreadPoolExecutor(1,TimeUnit.SECONDS,new LinkedBlockingQueue<Runnable>(),new NamedThreadFactory("MemtableFlushWriter"),"internal");

postFlushExecutor = new JMXEnabledThreadPoolExecutor(1,StageManager.KEEPALIVE,TimeUnit.SECONDS,new LinkedBlockingQueue<Runnable>(),new NamedThreadFactory("MemtablePostFlush"),"internal");

参考

https://wiki.apache.org/cassandra/WritePathForUsers

http://www.mikeperham.com/2010/03/13/cassandra-internals-writing/

号外

打算翻译cassandra官方文档,有兴趣的小伙伴可以一起加入
cassandra官方文档翻译
加入cassandra学习群,一起学习
从Stage角度看cassandra write