Examples with WorkerBackpressureCallback org.apache.storm.utils.WorkerBackpressureCallback used on opensource projects

Example 1 with WorkerBackpressureCallback

use of org.apache.storm.utils.WorkerBackpressureCallback in project storm by apache.

the class Worker method mkBackpressureHandler.

/**
     * make a handler that checks and updates worker's backpressure flag
     */
private WorkerBackpressureCallback mkBackpressureHandler() {
    final List<IRunningExecutor> executors = executorsAtom.get();
    return new WorkerBackpressureCallback() {

        @Override
        public void onEvent(Object obj) {
            String topologyId = workerState.topologyId;
            String assignmentId = workerState.assignmentId;
            int port = workerState.port;
            IStormClusterState stormClusterState = workerState.stormClusterState;
            boolean prevBackpressureFlag = workerState.backpressure.get();
            boolean currBackpressureFlag = prevBackpressureFlag;
            if (null != executors) {
                currBackpressureFlag = workerState.transferQueue.getThrottleOn() || (executors.stream().map(IRunningExecutor::getBackPressureFlag).reduce((op1, op2) -> (op1 || op2)).get());
            }
            if (currBackpressureFlag != prevBackpressureFlag) {
                try {
                    LOG.debug("worker backpressure flag changing from {} to {}", prevBackpressureFlag, currBackpressureFlag);
                    stormClusterState.workerBackpressure(topologyId, assignmentId, (long) port, currBackpressureFlag);
                    // doing the local reset after the zk update succeeds is very important to avoid a bad state upon zk exception
                    workerState.backpressure.set(currBackpressureFlag);
                } catch (Exception ex) {
                    LOG.error("workerBackpressure update failed when connecting to ZK ... will retry", ex);
                }
            }
        }
    };
}

Example 2 with WorkerBackpressureCallback

use of org.apache.storm.utils.WorkerBackpressureCallback in project storm by apache.

the class Worker method start.

public void start() throws Exception {
    LOG.info("Launching worker for {} on {}:{} with id {} and conf {}", topologyId, assignmentId, port, workerId, conf);
    // if ConfigUtils.isLocalMode(conf) returns false then it is in distributed mode.
    if (!ConfigUtils.isLocalMode(conf)) {
        // Distributed mode
        SysOutOverSLF4J.sendSystemOutAndErrToSLF4J();
        String pid = Utils.processPid();
        FileUtils.touch(new File(ConfigUtils.workerPidPath(conf, workerId, pid)));
        FileUtils.writeStringToFile(new File(ConfigUtils.workerArtifactsPidPath(conf, topologyId, port)), pid, Charset.forName("UTF-8"));
    }
    final Map topologyConf = ConfigUtils.overrideLoginConfigWithSystemProperty(ConfigUtils.readSupervisorStormConf(conf, topologyId));
    List<ACL> acls = Utils.getWorkerACL(topologyConf);
    IStateStorage stateStorage = ClusterUtils.mkStateStorage(conf, topologyConf, acls, new ClusterStateContext(DaemonType.WORKER));
    IStormClusterState stormClusterState = ClusterUtils.mkStormClusterState(stateStorage, acls, new ClusterStateContext());
    Credentials initialCredentials = stormClusterState.credentials(topologyId, null);
    Map<String, String> initCreds = new HashMap<>();
    if (initialCredentials != null) {
        initCreds.putAll(initialCredentials.get_creds());
    }
    autoCreds = AuthUtils.GetAutoCredentials(topologyConf);
    subject = AuthUtils.populateSubject(null, autoCreds, initCreds);
    Subject.doAs(subject, new PrivilegedExceptionAction<Object>() {

        @Override
        public Object run() throws Exception {
            workerState = new WorkerState(conf, context, topologyId, assignmentId, port, workerId, topologyConf, stateStorage, stormClusterState);
            // Heartbeat here so that worker process dies if this fails
            // it's important that worker heartbeat to supervisor ASAP so that supervisor knows
            // that worker is running and moves on
            doHeartBeat();
            executorsAtom = new AtomicReference<>(null);
            // launch heartbeat threads immediately so that slow-loading tasks don't cause the worker to timeout
            // to the supervisor
            workerState.heartbeatTimer.scheduleRecurring(0, (Integer) conf.get(Config.WORKER_HEARTBEAT_FREQUENCY_SECS), () -> {
                try {
                    doHeartBeat();
                } catch (IOException e) {
                    throw new RuntimeException(e);
                }
            });
            workerState.executorHeartbeatTimer.scheduleRecurring(0, (Integer) conf.get(Config.WORKER_HEARTBEAT_FREQUENCY_SECS), Worker.this::doExecutorHeartbeats);
            workerState.registerCallbacks();
            workerState.refreshConnections(null);
            workerState.activateWorkerWhenAllConnectionsReady();
            workerState.refreshStormActive(null);
            workerState.runWorkerStartHooks();
            List<IRunningExecutor> newExecutors = new ArrayList<IRunningExecutor>();
            for (List<Long> e : workerState.getExecutors()) {
                if (ConfigUtils.isLocalMode(topologyConf)) {
                    newExecutors.add(LocalExecutor.mkExecutor(workerState, e, initCreds).execute());
                } else {
                    newExecutors.add(Executor.mkExecutor(workerState, e, initCreds).execute());
                }
            }
            executorsAtom.set(newExecutors);
            EventHandler<Object> tupleHandler = (packets, seqId, batchEnd) -> workerState.sendTuplesToRemoteWorker((HashMap<Integer, ArrayList<TaskMessage>>) packets, seqId, batchEnd);
            // This thread will publish the messages destined for remote tasks to remote connections
            transferThread = Utils.asyncLoop(() -> {
                workerState.transferQueue.consumeBatchWhenAvailable(tupleHandler);
                return 0L;
            });
            DisruptorBackpressureCallback disruptorBackpressureHandler = mkDisruptorBackpressureHandler(workerState);
            workerState.transferQueue.registerBackpressureCallback(disruptorBackpressureHandler);
            workerState.transferQueue.setEnableBackpressure((Boolean) topologyConf.get(Config.TOPOLOGY_BACKPRESSURE_ENABLE));
            workerState.transferQueue.setHighWaterMark(Utils.getDouble(topologyConf.get(Config.BACKPRESSURE_DISRUPTOR_HIGH_WATERMARK)));
            workerState.transferQueue.setLowWaterMark(Utils.getDouble(topologyConf.get(Config.BACKPRESSURE_DISRUPTOR_LOW_WATERMARK)));
            WorkerBackpressureCallback backpressureCallback = mkBackpressureHandler();
            backpressureThread = new WorkerBackpressureThread(workerState.backpressureTrigger, workerState, backpressureCallback);
            if ((Boolean) topologyConf.get(Config.TOPOLOGY_BACKPRESSURE_ENABLE)) {
                backpressureThread.start();
                stormClusterState.topologyBackpressure(topologyId, workerState::refreshThrottle);
                int pollingSecs = Utils.getInt(topologyConf.get(Config.TASK_BACKPRESSURE_POLL_SECS));
                workerState.refreshBackpressureTimer.scheduleRecurring(0, pollingSecs, workerState::refreshThrottle);
            }
            credentialsAtom = new AtomicReference<Credentials>(initialCredentials);
            establishLogSettingCallback();
            workerState.stormClusterState.credentials(topologyId, Worker.this::checkCredentialsChanged);
            workerState.refreshCredentialsTimer.scheduleRecurring(0, (Integer) conf.get(Config.TASK_CREDENTIALS_POLL_SECS), new Runnable() {

                @Override
                public void run() {
                    checkCredentialsChanged();
                    if ((Boolean) topologyConf.get(Config.TOPOLOGY_BACKPRESSURE_ENABLE)) {
                        checkThrottleChanged();
                    }
                }
            });
            // The jitter allows the clients to get the data at different times, and avoids thundering herd
            if (!(Boolean) topologyConf.get(Config.TOPOLOGY_DISABLE_LOADAWARE_MESSAGING)) {
                workerState.refreshLoadTimer.scheduleRecurringWithJitter(0, 1, 500, workerState::refreshLoad);
            }
            workerState.refreshConnectionsTimer.scheduleRecurring(0, (Integer) conf.get(Config.TASK_REFRESH_POLL_SECS), workerState::refreshConnections);
            workerState.resetLogLevelsTimer.scheduleRecurring(0, (Integer) conf.get(Config.WORKER_LOG_LEVEL_RESET_POLL_SECS), logConfigManager::resetLogLevels);
            workerState.refreshActiveTimer.scheduleRecurring(0, (Integer) conf.get(Config.TASK_REFRESH_POLL_SECS), workerState::refreshStormActive);
            LOG.info("Worker has topology config {}", Utils.redactValue(topologyConf, Config.STORM_ZOOKEEPER_TOPOLOGY_AUTH_PAYLOAD));
            LOG.info("Worker {} for storm {} on {}:{}  has finished loading", workerId, topologyId, assignmentId, port);
            return this;
        }

        ;
    });
}