org.apache.hadoop.mapred.JvmManager Maven / Gradle / Ivy
The newest version!
/**
* 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.mapred;
import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Random;
import java.util.Vector;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.mapred.TaskController.TaskControllerContext;
import org.apache.hadoop.mapred.TaskTracker.TaskInProgress;
import org.apache.hadoop.mapreduce.TaskType;
import org.apache.hadoop.mapreduce.server.tasktracker.TTConfig;
import org.apache.hadoop.util.Shell.ShellCommandExecutor;
import org.apache.hadoop.util.StringUtils;
import org.apache.hadoop.mapreduce.util.ProcessTree;
class JvmManager {
public static final Log LOG =
LogFactory.getLog(JvmManager.class);
private JvmManagerForType mapJvmManager;
private JvmManagerForType reduceJvmManager;
public JvmEnv constructJvmEnv(List setup, Vectorvargs,
File stdout,File stderr,long logSize, File workDir,
Map env, JobConf conf) {
return new JvmEnv(setup,vargs,stdout,stderr,logSize,workDir,env,conf);
}
public JvmManager(TaskTracker tracker) {
mapJvmManager = new JvmManagerForType(tracker.getMaxCurrentMapTasks(),
true, tracker);
reduceJvmManager = new JvmManagerForType(tracker.getMaxCurrentReduceTasks(),
false, tracker);
}
JvmManagerForType getJvmManagerForType(TaskType type) {
if (type.equals(TaskType.MAP)) {
return mapJvmManager;
} else if (type.equals(TaskType.REDUCE)) {
return reduceJvmManager;
}
return null;
}
public void stop() {
mapJvmManager.stop();
reduceJvmManager.stop();
}
public boolean isJvmKnown(JVMId jvmId) {
if (jvmId.isMapJVM()) {
return mapJvmManager.isJvmknown(jvmId);
} else {
return reduceJvmManager.isJvmknown(jvmId);
}
}
/*
* Saves pid of the given taskJvm
*/
void setPidToJvm(JVMId jvmId, String pid) {
if (jvmId.isMapJVM()) {
mapJvmManager.setPidForJvm(jvmId, pid);
}
else {
reduceJvmManager.setPidForJvm(jvmId, pid);
}
}
/*
* Returns the pid of the task
*/
String getPid(TaskRunner t) {
if (t != null && t.getTask() != null) {
if (t.getTask().isMapTask()) {
return mapJvmManager.getPidByRunningTask(t);
} else {
return reduceJvmManager.getPidByRunningTask(t);
}
}
return null;
}
public void launchJvm(TaskRunner t, JvmEnv env) {
if (t.getTask().isMapTask()) {
mapJvmManager.reapJvm(t, env);
} else {
reduceJvmManager.reapJvm(t, env);
}
}
public TaskInProgress getTaskForJvm(JVMId jvmId)
throws IOException {
if (jvmId.isMapJVM()) {
return mapJvmManager.getTaskForJvm(jvmId);
} else {
return reduceJvmManager.getTaskForJvm(jvmId);
}
}
public void taskFinished(TaskRunner tr) {
if (tr.getTask().isMapTask()) {
mapJvmManager.taskFinished(tr);
} else {
reduceJvmManager.taskFinished(tr);
}
}
public void taskKilled(TaskRunner tr) {
if (tr.getTask().isMapTask()) {
mapJvmManager.taskKilled(tr);
} else {
reduceJvmManager.taskKilled(tr);
}
}
void dumpStack(TaskRunner tr) {
if (tr.getTask().isMapTask()) {
mapJvmManager.dumpStack(tr);
} else {
reduceJvmManager.dumpStack(tr);
}
}
public void killJvm(JVMId jvmId) {
if (jvmId.isMap) {
mapJvmManager.killJvm(jvmId);
} else {
reduceJvmManager.killJvm(jvmId);
}
}
/**
* Adds the task's work dir to the cleanup queue of taskTracker for
* asynchronous deletion of work dir.
* @param tracker taskTracker
* @param task the task whose work dir needs to be deleted
* @throws IOException
*/
static void deleteWorkDir(TaskTracker tracker, Task task) throws IOException {
tracker.getCleanupThread().addToQueue(
TaskTracker.buildTaskControllerTaskPathDeletionContexts(
tracker.getLocalFileSystem(),
tracker.getLocalFiles(tracker.getJobConf(), ""),
task, true /* workDir */,
tracker.getTaskController()));
}
static class JvmManagerForType {
//Mapping from the JVM IDs to running Tasks
Map jvmToRunningTask =
new HashMap();
//Mapping from the tasks to JVM IDs
Map runningTaskToJvm =
new HashMap();
//Mapping from the JVM IDs to Reduce JVM processes
Map jvmIdToRunner =
new HashMap();
int maxJvms;
boolean isMap;
TaskTracker tracker;
Random rand = new Random(System.currentTimeMillis());
public JvmManagerForType(int maxJvms, boolean isMap,
TaskTracker tracker) {
this.maxJvms = maxJvms;
this.isMap = isMap;
this.tracker = tracker;
}
synchronized public void setRunningTaskForJvm(JVMId jvmId,
TaskRunner t) {
jvmToRunningTask.put(jvmId, t);
runningTaskToJvm.put(t,jvmId);
jvmIdToRunner.get(jvmId).setTaskRunner(t);
}
synchronized public TaskInProgress getTaskForJvm(JVMId jvmId)
throws IOException {
if (jvmToRunningTask.containsKey(jvmId)) {
//Incase of JVM reuse, tasks are returned to previously launched
//JVM via this method. However when a new task is launched
//the task being returned has to be initialized.
TaskRunner taskRunner = jvmToRunningTask.get(jvmId);
JvmRunner jvmRunner = jvmIdToRunner.get(jvmId);
Task task = taskRunner.getTaskInProgress().getTask();
// Initialize task dirs
TaskControllerContext context =
new TaskController.TaskControllerContext();
context.env = jvmRunner.env;
context.task = task;
// If we are returning the same task as which the JVM was launched
// we don't initialize task once again.
if (!jvmRunner.env.conf.get(JobContext.TASK_ATTEMPT_ID).equals(
task.getTaskID().toString())) {
try {
tracker.getTaskController().initializeTask(context);
} catch (IOException e) {
LOG.warn("Failed to initialize the new task "
+ task.getTaskID().toString() + " to be given to JVM with id "
+ jvmId);
throw e;
}
}
return taskRunner.getTaskInProgress();
}
return null;
}
synchronized String getPidByRunningTask(TaskRunner t) {
JVMId id = runningTaskToJvm.get(t);
if (id != null) {
return jvmIdToRunner.get(id).getPid();
}
return null;
}
synchronized void setPidForJvm(JVMId jvmId, String pid) {
JvmRunner runner = jvmIdToRunner.get(jvmId);
assert runner != null : "Task must have a runner to set a pid";
runner.setPid(pid);
}
synchronized public boolean isJvmknown(JVMId jvmId) {
return jvmIdToRunner.containsKey(jvmId);
}
synchronized public void taskFinished(TaskRunner tr) {
JVMId jvmId = runningTaskToJvm.remove(tr);
if (jvmId != null) {
jvmToRunningTask.remove(jvmId);
JvmRunner jvmRunner;
if ((jvmRunner = jvmIdToRunner.get(jvmId)) != null) {
jvmRunner.taskRan();
}
}
}
synchronized public void taskKilled(TaskRunner tr) {
JVMId jvmId = runningTaskToJvm.remove(tr);
if (jvmId != null) {
jvmToRunningTask.remove(jvmId);
killJvm(jvmId);
}
}
synchronized public void killJvm(JVMId jvmId) {
JvmRunner jvmRunner;
if ((jvmRunner = jvmIdToRunner.get(jvmId)) != null) {
killJvmRunner(jvmRunner);
}
}
private synchronized void killJvmRunner(JvmRunner jvmRunner) {
jvmRunner.kill();
removeJvm(jvmRunner.jvmId);
}
void dumpStack(TaskRunner tr) {
JvmRunner jvmRunner = null;
synchronized (this) {
JVMId jvmId = runningTaskToJvm.get(tr);
if (null != jvmId) {
jvmRunner = jvmIdToRunner.get(jvmId);
}
}
// Don't want to hold JvmManager lock while dumping stacks for one
// task.
if (null != jvmRunner) {
jvmRunner.dumpChildStacks();
}
}
synchronized public void stop() {
//since the kill() method invoked later on would remove
//an entry from the jvmIdToRunner map, we create a
//copy of the values and iterate over it (if we don't
//make a copy, we will encounter concurrentModification
//exception
List list = new ArrayList();
list.addAll(jvmIdToRunner.values());
for (JvmRunner jvm : list) {
killJvmRunner(jvm);
}
}
synchronized private void removeJvm(JVMId jvmId) {
jvmIdToRunner.remove(jvmId);
}
private synchronized void reapJvm(
TaskRunner t, JvmEnv env) {
if (t.getTaskInProgress().wasKilled()) {
//the task was killed in-flight
//no need to do the rest of the operations
return;
}
boolean spawnNewJvm = false;
JobID jobId = t.getTask().getJobID();
//Check whether there is a free slot to start a new JVM.
//,or, Kill a (idle) JVM and launch a new one
//When this method is called, we *must*
// (1) spawn a new JVM (if we are below the max)
// (2) find an idle JVM (that belongs to the same job), or,
// (3) kill an idle JVM (from a different job)
// (the order of return is in the order above)
int numJvmsSpawned = jvmIdToRunner.size();
JvmRunner runnerToKill = null;
if (numJvmsSpawned >= maxJvms) {
//go through the list of JVMs for all jobs.
Iterator> jvmIter =
jvmIdToRunner.entrySet().iterator();
while (jvmIter.hasNext()) {
JvmRunner jvmRunner = jvmIter.next().getValue();
JobID jId = jvmRunner.jvmId.getJobId();
//look for a free JVM for this job; if one exists then just break
if (jId.equals(jobId) && !jvmRunner.isBusy() && !jvmRunner.ranAll()){
setRunningTaskForJvm(jvmRunner.jvmId, t); //reserve the JVM
LOG.info("No new JVM spawned for jobId/taskid: " +
jobId+"/"+t.getTask().getTaskID() +
". Attempting to reuse: " + jvmRunner.jvmId);
return;
}
//Cases when a JVM is killed:
// (1) the JVM under consideration belongs to the same job
// (passed in the argument). In this case, kill only when
// the JVM ran all the tasks it was scheduled to run (in terms
// of count).
// (2) the JVM under consideration belongs to a different job and is
// currently not busy
//But in both the above cases, we see if we can assign the current
//task to an idle JVM (hence we continue the loop even on a match)
if ((jId.equals(jobId) && jvmRunner.ranAll()) ||
(!jId.equals(jobId) && !jvmRunner.isBusy())) {
runnerToKill = jvmRunner;
spawnNewJvm = true;
}
}
} else {
spawnNewJvm = true;
}
if (spawnNewJvm) {
if (runnerToKill != null) {
LOG.info("Killing JVM: " + runnerToKill.jvmId);
killJvmRunner(runnerToKill);
}
spawnNewJvm(jobId, env, t);
return;
}
//*MUST* never reach this
throw new RuntimeException("Inconsistent state!!! " +
"JVM Manager reached an unstable state " +
"while reaping a JVM for task: " + t.getTask().getTaskID()+
" " + getDetails());
}
private synchronized String getDetails() {
StringBuffer details = new StringBuffer();
details.append("Number of active JVMs:").
append(jvmIdToRunner.size());
Iterator jvmIter =
jvmIdToRunner.keySet().iterator();
while (jvmIter.hasNext()) {
JVMId jvmId = jvmIter.next();
details.append("\n JVMId ").
append(jvmId.toString()).
append(" #Tasks ran: ").
append(jvmIdToRunner.get(jvmId).numTasksRan).
append(" Currently busy? ").
append(jvmIdToRunner.get(jvmId).isBusy()).
append(" Currently running: ").
append(jvmToRunningTask.get(jvmId).getTask().getTaskID().toString());
}
return details.toString();
}
private synchronized void spawnNewJvm(JobID jobId, JvmEnv env,
TaskRunner t) {
JvmRunner jvmRunner = new JvmRunner(env,jobId);
jvmIdToRunner.put(jvmRunner.jvmId, jvmRunner);
//spawn the JVM in a new thread. Note that there will be very little
//extra overhead of launching the new thread for a new JVM since
//most of the cost is involved in launching the process. Moreover,
//since we are going to be using the JVM for running many tasks,
//the thread launch cost becomes trivial when amortized over all
//tasks. Doing it this way also keeps code simple.
jvmRunner.setDaemon(true);
jvmRunner.setName("JVM Runner " + jvmRunner.jvmId + " spawned.");
setRunningTaskForJvm(jvmRunner.jvmId, t);
LOG.info(jvmRunner.getName());
jvmRunner.start();
}
synchronized private void updateOnJvmExit(JVMId jvmId,
int exitCode) {
removeJvm(jvmId);
TaskRunner t = jvmToRunningTask.remove(jvmId);
if (t != null) {
runningTaskToJvm.remove(t);
if (exitCode != 0) {
t.setExitCode(exitCode);
}
t.signalDone();
}
}
class JvmRunner extends Thread {
JvmEnv env;
volatile boolean killed = false;
volatile int numTasksRan;
final int numTasksToRun;
JVMId jvmId;
private ShellCommandExecutor shexec; // shell terminal for running the task
//context used for starting JVM
private TaskControllerContext initalContext;
public JvmRunner(JvmEnv env, JobID jobId) {
this.env = env;
this.jvmId = new JVMId(jobId, isMap, rand.nextInt());
this.numTasksToRun = env.conf.getNumTasksToExecutePerJvm();
this.initalContext = new TaskControllerContext();
initalContext.sleeptimeBeforeSigkill = tracker.getJobConf()
.getLong(TTConfig.TT_SLEEP_TIME_BEFORE_SIG_KILL,
ProcessTree.DEFAULT_SLEEPTIME_BEFORE_SIGKILL);
LOG.info("In JvmRunner constructed JVM ID: " + jvmId);
}
public void run() {
runChild(env);
}
public void runChild(JvmEnv env) {
try {
env.vargs.add(Integer.toString(jvmId.getId()));
//Launch the task controller to run task JVM
initalContext.env = env;
tracker.getTaskController().launchTaskJVM(initalContext);
} catch (IOException ioe) {
// do nothing
// error and output are appropriately redirected
} finally { // handle the exit code
shexec = initalContext.shExec;
if (shexec == null) {
return;
}
kill();
int exitCode = shexec.getExitCode();
updateOnJvmExit(jvmId, exitCode);
LOG.info("JVM : " + jvmId + " exited with exit code " + exitCode
+ ". Number of tasks it ran: " + numTasksRan);
try {
// In case of jvm-reuse,
//the task jvm cleans up the common workdir for every
//task at the beginning of each task in the task JVM.
//For the last task, we do it here.
if (env.conf.getNumTasksToExecutePerJvm() != 1) {
deleteWorkDir(tracker, initalContext.task);
}
} catch (IOException ie){}
}
}
synchronized void setPid(String pid) {
assert initalContext != null;
initalContext.pid = pid;
}
synchronized String getPid() {
if (initalContext != null) {
return initalContext.pid;
} else {
return null;
}
}
/**
* Kills the process. Also kills its subprocesses if the process(root of subtree
* of processes) is created using setsid.
*/
synchronized void kill() {
if (!killed) {
TaskController controller = tracker.getTaskController();
// Check inital context before issuing a kill to prevent situations
// where kill is issued before task is launched.
if (initalContext != null && initalContext.env != null) {
// Destroy the task jvm
controller.destroyTaskJVM(initalContext);
} else {
LOG.info(String.format("JVM Not killed %s but just removed", jvmId
.toString()));
}
killed = true;
}
}
/** Send a signal to the JVM requesting that it dump a stack trace,
* and wait for a timeout interval to give this signal time to be
* processed.
*/
void dumpChildStacks() {
if (!killed) {
TaskController controller = tracker.getTaskController();
// Check inital context before issuing a signal to prevent situations
// where signal is issued before task is launched.
if (initalContext != null && initalContext.env != null) {
// signal the task jvm
controller.dumpTaskStack(initalContext);
// We're going to kill the jvm with SIGKILL after this,
// so we should wait for a few seconds first to ensure that
// the SIGQUIT has time to be processed.
try {
Thread.sleep(initalContext.sleeptimeBeforeSigkill);
} catch (InterruptedException e) {
LOG.warn("Sleep interrupted : " +
StringUtils.stringifyException(e));
}
}
}
}
public synchronized void taskRan() {
initalContext.task = null;
numTasksRan++;
}
public boolean ranAll() {
return(numTasksRan == numTasksToRun);
}
public synchronized void setTaskRunner(TaskRunner runner) {
initalContext.task = runner.getTask();
assert initalContext.task != null;
}
public synchronized boolean isBusy() {
return initalContext.task != null;
}
}
}
static class JvmEnv { //Helper class
List vargs;
List setup;
File stdout;
File stderr;
File workDir;
long logSize;
JobConf conf;
Map env;
public JvmEnv(List setup, Vector vargs, File stdout,
File stderr, long logSize, File workDir, Map env,
JobConf conf) {
this.setup = setup;
this.vargs = vargs;
this.stdout = stdout;
this.stderr = stderr;
this.workDir = workDir;
this.env = env;
this.conf = conf;
}
}
}