Twitter Storm中Bolt消息传递路径之源码解读

本文初次发表于storm-cn的google groups中，现以blog的方式再次发表，表明本人徽沪一郎确实读过这些代码，:).

Bolt作为task被executor执行，而executor是一个个的线程，所以executor必须存在于具体的process之中，而这个process就是worker。至于worker是如何被supervisor创建，尔后worker又如何创建executor线程，这些暂且按下不表。

假设同属于一个Topology的Spout与Bolt分别处于不同的JVM，即不同的worker中，不同的JVM可能处于同一台物理机器，也可能处于不同的物理机器中。为了让情景简单，认为JVM处于不同的物理机器中。

Spout的输出消息到达Bolt，作为Bolt的输入会经过这么几个阶段。

1. spout的输出通过该spout所处worker的消息输出线程，将tuple输入到Bolt所属的worker。它们之间的通路是socket连接，用ZeroMQ实现。
2. bolt所处的worker有一个专门处理socket消息的receive thread 接收到spout发送来的tuple
3. receive thread将接收到的消息传送给对应的bolt所在的executor。 在worker内部（即同一process内部),消息传递使用的是Lmax Disruptor pattern.
4. executor接收到tuple之后，由event-handler进行处理

下面是具体的源码
1. worker创建消息接收线程

worker.clj

(defn launch-receive-thread [worker]
(log-message "Launching receive-thread for " (:assignment-id worker) ":" (:port worker))
(msg-loader/launch-receive-thread!
(:mq-context worker)
(:storm-id worker)
(:port worker)
(:transfer-local-fn worker)
(-> worker :storm-conf (get TOPOLOGY-RECEIVER-BUFFER-SIZE))
:kill-fn (fn [t] (halt-process! 11))))

注意加亮的行会将storm.yaml中配置使用ZMQ或其它
storm.messaging.transport:"backtype.storm.messaging.zmq"

2. worker从socket接收到新消息

vthread (async-loop
(fn []
(let [socket (.bind ^IContext context storm-id port)]
(fn []
(let [batched (ArrayList.)
init (.recv ^IConnection socket 0)]
(loop [packet init]
(let [task (if packet (.task ^TaskMessage packet))
message (if packet (.message ^TaskMessage packet))]
(if (= task -1)
(do (log-message "Receiving-thread:[" storm-id ", " port "] received shutdown notice")
(.close socket)
nil )
(do
(when packet (.add batched [task message]))
(if (and packet (< (.size batched) max-buffer-size))
(recur (.recv ^IConnection socket 1))
(do (transfer-local-fn batched)
0 ))))))))))

加亮行使用的transfer-local-fn会将接收的TaskMessage传递给相应的executor

3. transfer-local-fn

(defn mk-transfer-local-fn [worker]
(let [short-executor-receive-queue-map (:short-executor-receive-queue-map worker)
task->short-executor (:task->short-executor worker)
task-getter (comp #(get task->short-executor %) fast-first)]
(fn [tuple-batch]
(let [grouped (fast-group-by task-getter tuple-batch)]
(fast-map-iter [[short-executor pairs] grouped]
(let [q (short-executor-receive-queue-map short-executor)]
(if q
(disruptor/publish q pairs)
(log-warn "Received invalid messages for unknown tasks. Dropping... ")
)))))))

用disruptor在线程之间进行消息传递。

多费一句话，mk-transfer-local-fn表示将外部世界的消息传递给本进程内的线程。而mk-transfer-fn则刚好在方向上反过来。

4. 消息被executor处理

executor.clj
==========================================================

(defn mk-task-receiver [executor-data tuple-action-fn]
(let [^KryoTupleDeserializer deserializer (:deserializer executor-data)
task-ids (:task-ids executor-data)
debug? (= true (-> executor-data :storm-conf (get TOPOLOGY-DEBUG)))
]
(disruptor/clojure-handler
(fn [tuple-batch sequence-id end-of-batch?]
(fast-list-iter [[task-id msg] tuple-batch]
(let [^TupleImpl tuple (if (instance? Tuple msg) msg (.deserialize deserializer msg))]
(when debug? (log-message "Processing received message " tuple))
(if task-id
(tuple-action-fn task-id tuple)
;; null task ids are broadcast tuples
(fast-list-iter [task-id task-ids]
(tuple-action-fn task-id tuple)
))
))))))

加亮行中tuple-action-fn定义于mk-threads(源文件executor.clj)中。因为当前以Bolt为例，所以会调用的tuple-action-fn定义于defmethod mk-threads :bolt [executor-data task-datas]

那么mk-task-receiver是如何与disruptor关联起来的呢，可以见定义于mk-threads中的下述代码

(let [receive-queue (:receive-queue executor-data)
event-handler (mk-task-receiver executor-data tuple-action-fn)]
(disruptor/consumer-started! receive-queue)
(fn []
(disruptor/consume-batch-when-available receive-queue event-handler)
0)))

到了这里，消息的发送与接收处理路径打通。