【问题标题】:worker disconnecting with ZK for long running bolt工人与 ZK 断开连接以获取长时间运行的螺栓
【发布时间】:2016-09-09 09:06:31
【问题描述】:

我有一个使用大型对象池的螺栓。当池重新初始化(每 4 小时一次)时,bolt 等待几秒钟并与 zookeper 断开连接。 我在 yaml 中指定了以下属性,但工人仍然死亡。

supervisor.worker.start.timeout.secs 300
supervisor.worker.timeout.secs 60

这里是日志:

2015-02-17 04:35:01 o.a.z.ClientCnxn [INFO] Client session timed out, have not heard from server in 15906ms for sessionid 0x14b9200ea400009, closing socket connection and attempting reconnect
2015-02-17 04:35:01 o.a.c.f.s.ConnectionStateManager [INFO] State change: SUSPENDED
2015-02-17 04:35:01 o.a.c.f.s.ConnectionStateManager [WARN] There are no ConnectionStateListeners registered.
2015-02-17 04:35:01 b.s.cluster [WARN] Received event :disconnected::none: with disconnected Zookeeper.
2015-02-17 04:35:02 o.a.z.ClientCnxn [INFO] Opening socket connection to server realtimeanalytics.novalocal/10.0.0.11:2181. Will not attempt to authenticate using SASL (Unable to locate a login configuration)
2015-02-17 04:35:02 o.a.z.ClientCnxn [INFO] Socket connection established to realtimeanalytics.novalocal/10.0.0.11:2181, initiating session
2015-02-17 04:35:02 o.a.z.ClientCnxn [INFO] Session establishment complete on server realtimeanalytics.novalocal/10.0.0.11:2181, sessionid = 0x14b9200ea400009, negotiated timeout = 20000
2015-02-17 04:35:02 o.a.c.f.s.ConnectionStateManager [INFO] State change: RECONNECTED
2015-02-17 04:35:02 o.a.c.f.s.ConnectionStateManager [WARN] There are no ConnectionStateListeners registered.
2015-02-17 04:35:33 o.a.z.ClientCnxn [INFO] Client session timed out, have not heard from server in 13499ms for sessionid 0x14b9200ea400009, closing socket connection and attempting reconnect
2015-02-17 04:35:34 o.a.c.f.s.ConnectionStateManager [INFO] State change: SUSPENDED
2015-02-17 04:35:34 b.s.cluster [WARN] Received event :disconnected::none: with disconnected Zookeeper.
2015-02-17 04:35:34 o.a.c.f.s.ConnectionStateManager [WARN] There are no ConnectionStateListeners registered.

【问题讨论】:

    标签: apache-storm apache-zookeeper


    【解决方案1】:

    您是否尝试过增加工作器心跳超时?

    worker.heartbeat.frequency.secs: 60
    

    【讨论】:

    • zookeeper 问题已通过添加storm.zookeeper.session.timeout: 40000 storm.zookeeper.connection.timeout: 30000 解决但仍将工作人员移至其他节点可能是因为内存压力
    猜你喜欢
    • 1970-01-01
    • 1970-01-01
    • 1970-01-01
    • 2019-04-24
    • 2016-10-30
    • 1970-01-01
    • 1970-01-01
    • 2014-09-02
    • 1970-01-01
    相关资源
    最近更新 更多