Package org.apache.hadoop.hbase.procedure

Source Code of org.apache.hadoop.hbase.procedure.ProcedureMember

/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements.  See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership.  The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License.  You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.hbase.procedure;

import java.io.Closeable;
import java.io.IOException;
import java.util.Collection;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Future;
import java.util.concurrent.RejectedExecutionException;
import java.util.concurrent.SynchronousQueue;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.hbase.DaemonThreadFactory;
import org.apache.hadoop.hbase.errorhandling.ForeignException;

import com.google.common.collect.MapMaker;

/**
* Process to kick off and manage a running {@link Subprocedure} on a member. This is the
* specialized part of a {@link Procedure} that actually does procedure type-specific work
* and reports back to the coordinator as it completes each phase.
* <p>
* If there is a connection error ({@link #controllerConnectionFailure(String, IOException)}), all
* currently running subprocedures are notify to failed since there is no longer a way to reach any
* other members or coordinators since the rpcs are down.
*/
@InterfaceAudience.Public
@InterfaceStability.Evolving
public class ProcedureMember implements Closeable {
  private static final Log LOG = LogFactory.getLog(ProcedureMember.class);

  final static long KEEP_ALIVE_MILLIS_DEFAULT = 5000;

  private final SubprocedureFactory builder;
  private final ProcedureMemberRpcs rpcs;

  private final ConcurrentMap<String,Subprocedure> subprocs =
      new MapMaker().concurrencyLevel(4).weakValues().makeMap();
  private final ExecutorService pool;

  /**
   * Instantiate a new ProcedureMember.  This is a slave that executes subprocedures.
   *
   * @param rpcs controller used to send notifications to the procedure coordinator
   * @param pool thread pool to submit subprocedures
   * @param factory class that creates instances of a subprocedure.
   */
  public ProcedureMember(ProcedureMemberRpcs rpcs, ThreadPoolExecutor pool,
      SubprocedureFactory factory) {
    this.pool = pool;
    this.rpcs = rpcs;
    this.builder = factory;
  }

  /**
   * Default thread pool for the procedure
   *
   * @param memberName
   * @param procThreads the maximum number of threads to allow in the pool
   */
  public static ThreadPoolExecutor defaultPool(String memberName, int procThreads) {
    return defaultPool(memberName, procThreads, KEEP_ALIVE_MILLIS_DEFAULT);
  }

  /**
   * Default thread pool for the procedure
   *
   * @param memberName
   * @param procThreads the maximum number of threads to allow in the pool
   * @param keepAliveMillis the maximum time (ms) that excess idle threads will wait for new tasks
   */
  public static ThreadPoolExecutor defaultPool(String memberName, int procThreads,
      long keepAliveMillis) {
    return new ThreadPoolExecutor(1, procThreads, keepAliveMillis, TimeUnit.MILLISECONDS,
        new SynchronousQueue<Runnable>(),
        new DaemonThreadFactory("member: '" + memberName + "' subprocedure-pool"));
  }

  /**
   * Package exposed.  Not for public use.
   *
   * @return reference to the Procedure member's rpcs object
   */
  ProcedureMemberRpcs getRpcs() {
    return rpcs;
  }


  /**
   * This is separated from execution so that we can detect and handle the case where the
   * subprocedure is invalid and inactionable due to bad info (like DISABLED snapshot type being
   * sent here)
   * @param opName
   * @param data
   * @return subprocedure
   */
  public Subprocedure createSubprocedure(String opName, byte[] data) {
    return builder.buildSubprocedure(opName, data);
  }

  /**
   * Submit an subprocedure for execution.  This starts the local acquire phase.
   * @param subproc the subprocedure to execute.
   * @return <tt>true</tt> if the subprocedure was started correctly, <tt>false</tt> if it
   *         could not be started. In the latter case, the subprocedure holds a reference to
   *         the exception that caused the failure.
   */
  public boolean submitSubprocedure(Subprocedure subproc) {
     // if the submitted subprocedure was null, bail.
    if (subproc == null) {
      LOG.warn("Submitted null subprocedure, nothing to run here.");
      return false;
    }

    String procName = subproc.getName();
    if (procName == null || procName.length() == 0) {
      LOG.error("Subproc name cannot be null or the empty string");
      return false;
    }

    // make sure we aren't already running an subprocedure of that name
    Subprocedure rsub;
    synchronized (subprocs) {
      rsub = subprocs.get(procName);
    }
    if (rsub != null) {
      if (!rsub.isComplete()) {
        LOG.error("Subproc '" + procName + "' is already running. Bailing out");
        return false;
      }
      LOG.warn("A completed old subproc "  +  procName + " is still present, removing");
      subprocs.remove(procName);
    }

    LOG.debug("Submitting new Subprocedure:" + procName);

    // kick off the subprocedure
    Future<Void> future = null;
    try {
      future = this.pool.submit(subproc);
      synchronized (subprocs) {
        subprocs.put(procName, subproc);
      }
      return true;
    } catch (RejectedExecutionException e) {
      // the thread pool is full and we can't run the subprocedure
      String msg = "Subprocedure pool is full!";
      subproc.cancel(msg, e.getCause());

      // cancel all subprocedures proactively
      if (future != null) {
        future.cancel(true);
      }
    }

    LOG.error("Failed to start subprocedure '" + procName + "'");
    return false;
  }

   /**
    * Notification that procedure coordinator has reached the global barrier
    * @param procName name of the subprocedure that should start running the the in-barrier phase
    */
   public void receivedReachedGlobalBarrier(String procName) {
     Subprocedure subproc = subprocs.get(procName);
     if (subproc == null) {
       LOG.warn("Unexpected reached glabal barrier message for Sub-Procedure '" + procName + "'");
       return;
     }
     subproc.receiveReachedGlobalBarrier();
   }

  /**
   * Best effort attempt to close the threadpool via Thread.interrupt.
   */
  @Override
  public void close() throws IOException {
    // have to use shutdown now to break any latch waiting
    pool.shutdownNow();
  }

  /**
   * Shutdown the threadpool, and wait for upto timeoutMs millis before bailing
   * @param timeoutMs timeout limit in millis
   * @return true if successfully, false if bailed due to timeout.
   * @throws InterruptedException
   */
  boolean closeAndWait(long timeoutMs) throws InterruptedException {
    pool.shutdown();
    return pool.awaitTermination(timeoutMs, TimeUnit.MILLISECONDS);
  }

  /**
   * The connection to the rest of the procedure group (member and coordinator) has been
   * broken/lost/failed. This should fail any interested subprocedure, but not attempt to notify
   * other members since we cannot reach them anymore.
   * @param message description of the error
   * @param cause the actual cause of the failure
   *
   * TODO i'm tempted to just remove this code completely and treat it like any other abort.
   * Implementation wise, if this happens it is a ZK failure which means the RS will abort.
   */
  public void controllerConnectionFailure(final String message, final IOException cause) {
    Collection<Subprocedure> toNotify = subprocs.values();
    LOG.error(message, cause);
    for (Subprocedure sub : toNotify) {
      // TODO notify the elements, if they aren't null
      sub.cancel(message, cause);
    }
  }

  /**
   * Send abort to the specified procedure
   * @param procName name of the procedure to about
   * @param ee exception information about the abort
   */
  public void receiveAbortProcedure(String procName, ForeignException ee) {
    LOG.debug("Request received to abort procedure " + procName, ee);
    // if we know about the procedure, notify it
    Subprocedure sub = subprocs.get(procName);
    if (sub == null) {
      LOG.info("Received abort on procedure with no local subprocedure " + procName +
          ", ignoring it.", ee);
      return; // Procedure has already completed
    }
    LOG.error("Propagating foreign exception to subprocedure " + sub.getName(), ee);
    sub.monitor.receive(ee);
  }
}
TOP

Related Classes of org.apache.hadoop.hbase.procedure.ProcedureMember

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.