/**
* We lost the task tracker! All task-tracker structures have already been updated. Just process
* the contained tasks and any jobs that might be affected.
*/
void lostTaskTracker(String trackerName) {
LOG.info("Lost tracker '" + trackerName + "'");
TreeSet lostTasks = (TreeSet) trackerToTaskMap.get(trackerName);
trackerToTaskMap.remove(trackerName);
if (lostTasks != null) {
for (Iterator it = lostTasks.iterator(); it.hasNext(); ) {
String taskId = (String) it.next();
TaskInProgress tip = (TaskInProgress) taskidToTIPMap.get(taskId);
// Tell the job to fail the relevant task
JobInProgress job = tip.getJob();
job.failedTask(tip, taskId, trackerName);
}
}
}
/**
* The run method lives for the life of the JobTracker, and removes Jobs that are not still
* running, but which finished a long time ago.
*/
public void run() {
while (shouldRun) {
try {
Thread.sleep(RETIRE_JOB_CHECK_INTERVAL);
} catch (InterruptedException ie) {
}
synchronized (jobs) {
synchronized (jobInitQueue) {
synchronized (jobsByArrival) {
for (Iterator it = jobs.keySet().iterator(); it.hasNext(); ) {
String jobid = (String) it.next();
JobInProgress job = (JobInProgress) jobs.get(jobid);
if (job.getStatus().getRunState() != JobStatus.RUNNING
&& job.getStatus().getRunState() != JobStatus.PREP
&& (job.getFinishTime() + RETIRE_JOB_INTERVAL < System.currentTimeMillis())) {
it.remove();
jobInitQueue.remove(job);
jobsByArrival.remove(job);
}
}
}
}
}
}
}
public void reduce(
DoubleWritable key,
Iterator<DoubleWritable> values,
OutputCollector<DoubleWritable, Text> output,
Reporter reporter)
throws IOException {
output.collect(key, new Text(values.next().toString() + " - "));
}
@Override
public void reduce(
IntWritable key,
Iterator<DoubleWritable> values,
OutputCollector<IntWritable, DoubleWritable> output,
Reporter reporter)
throws IOException {
output.collect(key, values.next());
}
public void reduce(
Text key,
Iterator<CrawlDatum> values,
OutputCollector<Text, CrawlDatum> output,
Reporter reporter)
throws IOException {
boolean oldSet = false;
boolean injectedSet = false;
while (values.hasNext()) {
CrawlDatum val = values.next();
if (val.getStatus() == CrawlDatum.STATUS_INJECTED) {
injected.set(val);
injected.setStatus(CrawlDatum.STATUS_DB_UNFETCHED);
injectedSet = true;
} else {
old.set(val);
oldSet = true;
}
}
CrawlDatum res = null;
/**
* Whether to overwrite, ignore or update existing records
*
* @see https://issues.apache.org/jira/browse/NUTCH-1405
*/
// Injected record already exists and overwrite but not update
if (injectedSet && oldSet && overwrite) {
res = injected;
if (update) {
LOG.info(key.toString() + " overwritten with injected record but update was specified.");
}
}
// Injected record already exists and update but not overwrite
if (injectedSet && oldSet && update && !overwrite) {
res = old;
old.putAllMetaData(injected);
old.setScore(injected.getScore() != scoreInjected ? injected.getScore() : old.getScore());
old.setFetchInterval(
injected.getFetchInterval() != interval
? injected.getFetchInterval()
: old.getFetchInterval());
}
// Old default behaviour
if (injectedSet && !oldSet) {
res = injected;
} else {
res = old;
}
output.collect(key, res);
}
public synchronized TaskReport[] getReduceTaskReports(String jobid) {
JobInProgress job = (JobInProgress) jobs.get(jobid);
if (job == null) {
return new TaskReport[0];
} else {
Vector reports = new Vector();
Vector completeReduceTasks = job.reportTasksInProgress(false, true);
for (Iterator it = completeReduceTasks.iterator(); it.hasNext(); ) {
TaskInProgress tip = (TaskInProgress) it.next();
reports.add(tip.generateSingleReport());
}
Vector incompleteReduceTasks = job.reportTasksInProgress(false, false);
for (Iterator it = incompleteReduceTasks.iterator(); it.hasNext(); ) {
TaskInProgress tip = (TaskInProgress) it.next();
reports.add(tip.generateSingleReport());
}
return (TaskReport[]) reports.toArray(new TaskReport[reports.size()]);
}
}
/** Return a table of block data */
public Block[] getBlockReport() {
TreeSet<Block> blockSet = new TreeSet<Block>();
volumes.getBlockInfo(blockSet);
Block blockTable[] = new Block[blockSet.size()];
int i = 0;
for (Iterator<Block> it = blockSet.iterator(); it.hasNext(); i++) {
blockTable[i] = it.next();
}
return blockTable;
}
public Vector runningJobs() {
Vector v = new Vector();
for (Iterator it = jobs.values().iterator(); it.hasNext(); ) {
JobInProgress jip = (JobInProgress) it.next();
JobStatus status = jip.getStatus();
if (status.getRunState() == JobStatus.RUNNING) {
v.add(jip);
}
}
return v;
}
public Vector completedJobs() {
Vector v = new Vector();
for (Iterator it = jobs.values().iterator(); it.hasNext(); ) {
JobInProgress jip = (JobInProgress) it.next();
JobStatus status = jip.getStatus();
if (status.getRunState() == JobStatus.SUCCEEDED) {
v.add(jip);
}
}
return v;
}
public void reduce(
Text key,
Iterator<IntWritable> values,
OutputCollector<Text, IntWritable> output,
Reporter reporter)
throws IOException {
int sum = 0;
while (values.hasNext()) {
sum += values.next().get();
}
output.collect(key, new IntWritable(sum));
}
/** Start the JobTracker process, listen on the indicated port */
JobTracker(Configuration conf) throws IOException {
//
// Grab some static constants
//
maxCurrentTasks = conf.getInt("mapred.tasktracker.tasks.maximum", 2);
RETIRE_JOB_INTERVAL = conf.getLong("mapred.jobtracker.retirejob.interval", 24 * 60 * 60 * 1000);
RETIRE_JOB_CHECK_INTERVAL = conf.getLong("mapred.jobtracker.retirejob.check", 60 * 1000);
TASK_ALLOC_EPSILON = conf.getFloat("mapred.jobtracker.taskalloc.loadbalance.epsilon", 0.2f);
PAD_FRACTION = conf.getFloat("mapred.jobtracker.taskalloc.capacitypad", 0.1f);
MIN_SLOTS_FOR_PADDING = 3 * maxCurrentTasks;
// This is a directory of temporary submission files. We delete it
// on startup, and can delete any files that we're done with
this.conf = conf;
JobConf jobConf = new JobConf(conf);
this.systemDir = jobConf.getSystemDir();
this.fs = FileSystem.get(conf);
FileUtil.fullyDelete(fs, systemDir);
fs.mkdirs(systemDir);
// Same with 'localDir' except it's always on the local disk.
jobConf.deleteLocalFiles(SUBDIR);
// Set ports, start RPC servers, etc.
InetSocketAddress addr = getAddress(conf);
this.localMachine = addr.getHostName();
this.port = addr.getPort();
this.interTrackerServer = RPC.getServer(this, addr.getPort(), 10, false, conf);
this.interTrackerServer.start();
Properties p = System.getProperties();
for (Iterator it = p.keySet().iterator(); it.hasNext(); ) {
String key = (String) it.next();
String val = (String) p.getProperty(key);
LOG.info("Property '" + key + "' is " + val);
}
this.infoPort = conf.getInt("mapred.job.tracker.info.port", 50030);
this.infoServer = new JobTrackerInfoServer(this, infoPort);
this.infoServer.start();
this.startTime = System.currentTimeMillis();
new Thread(this.expireTrackers).start();
new Thread(this.retireJobs).start();
new Thread(this.initJobs).start();
}
/**
* A tracker wants to know if any of its Tasks have been closed (because the job completed,
* whether successfully or not)
*/
public synchronized String pollForTaskWithClosedJob(String taskTracker) {
TreeSet taskIds = (TreeSet) trackerToTaskMap.get(taskTracker);
if (taskIds != null) {
for (Iterator it = taskIds.iterator(); it.hasNext(); ) {
String taskId = (String) it.next();
TaskInProgress tip = (TaskInProgress) taskidToTIPMap.get(taskId);
if (tip.shouldCloseForClosedJob(taskId)) {
//
// This is how the JobTracker ends a task at the TaskTracker.
// It may be successfully completed, or may be killed in
// mid-execution.
//
return taskId;
}
}
}
return null;
}
/**
* Accept and process a new TaskTracker profile. We might have known about the TaskTracker
* previously, or it might be brand-new. All task-tracker structures have already been updated.
* Just process the contained tasks and any jobs that might be affected.
*/
void updateTaskStatuses(TaskTrackerStatus status) {
for (Iterator it = status.taskReports(); it.hasNext(); ) {
TaskStatus report = (TaskStatus) it.next();
TaskInProgress tip = (TaskInProgress) taskidToTIPMap.get(report.getTaskId());
if (tip == null) {
LOG.info(
"Serious problem. While updating status, cannot find taskid " + report.getTaskId());
} else {
JobInProgress job = tip.getJob();
job.updateTaskStatus(tip, report);
if (report.getRunState() == TaskStatus.SUCCEEDED) {
job.completedTask(tip, report.getTaskId());
} else if (report.getRunState() == TaskStatus.FAILED) {
// Tell the job to fail the relevant task
job.failedTask(tip, report.getTaskId(), status.getTrackerName());
}
}
}
}
private static void deleteCache(Configuration conf, MRAsyncDiskService asyncDiskService)
throws IOException {
List<CacheStatus> deleteSet = new LinkedList<CacheStatus>();
// try deleting cache Status with refcount of zero
synchronized (cachedArchives) {
for (Iterator<String> it = cachedArchives.keySet().iterator(); it.hasNext(); ) {
String cacheId = (String) it.next();
CacheStatus lcacheStatus = cachedArchives.get(cacheId);
if (lcacheStatus.refcount == 0) {
// delete this cache entry from the global list
// and mark the localized file for deletion
deleteSet.add(lcacheStatus);
it.remove();
}
}
}
// do the deletion asynchronously, after releasing the global lock
Thread cacheFileCleaner =
new Thread(new CacheFileCleanTask(asyncDiskService, FileSystem.getLocal(conf), deleteSet));
cacheFileCleaner.start();
}
public void reduce(
IntWritable key,
Iterator<Text> values,
OutputCollector<IntWritable, Text> output,
Reporter reporter)
throws IOException {
HashMap<String, Integer> countries_map = new HashMap<String, Integer>();
ArrayList<Integer> counties = new ArrayList<>();
String cp = new String();
while (values.hasNext()) {
cp = values.next().toString();
if (countries_map.containsKey(cp)) {
countries_map.put(cp, countries_map.get(cp) + 1);
} else {
countries_map.put(cp, 1);
}
}
for (java.util.Map.Entry<String, Integer> entry : countries_map.entrySet()) {
counties.add(entry.getValue());
}
output.collect(
key,
new Text(
""
+ countries_map.entrySet().size()
+ " "
+ Collections.min(counties)
+ " "
+ median(counties)
+ " "
+ Collections.max(counties)
+ " "
+ mean(counties)
+ " "
+ standard_deviation(counties)));
}
/**
* Try to update an old block to a new block. If there are ongoing create threads running for the
* old block, the threads will be returned without updating the block.
*
* @return ongoing create threads if there is any. Otherwise, return null.
*/
private synchronized List<Thread> tryUpdateBlock(Block oldblock, Block newblock)
throws IOException {
// check ongoing create threads
final ActiveFile activefile = ongoingCreates.get(oldblock);
if (activefile != null && !activefile.threads.isEmpty()) {
// remove dead threads
for (Iterator<Thread> i = activefile.threads.iterator(); i.hasNext(); ) {
final Thread t = i.next();
if (!t.isAlive()) {
i.remove();
}
}
// return living threads
if (!activefile.threads.isEmpty()) {
return new ArrayList<Thread>(activefile.threads);
}
}
// No ongoing create threads is alive. Update block.
File blockFile = findBlockFile(oldblock.getBlockId());
if (blockFile == null) {
throw new IOException("Block " + oldblock + " does not exist.");
}
File oldMetaFile = findMetaFile(blockFile);
long oldgs = parseGenerationStamp(blockFile, oldMetaFile);
// rename meta file to a tmp file
File tmpMetaFile =
new File(
oldMetaFile.getParent(),
oldMetaFile.getName() + "_tmp" + newblock.getGenerationStamp());
if (!oldMetaFile.renameTo(tmpMetaFile)) {
throw new IOException("Cannot rename block meta file to " + tmpMetaFile);
}
// update generation stamp
if (oldgs > newblock.getGenerationStamp()) {
throw new IOException(
"Cannot update block (id="
+ newblock.getBlockId()
+ ") generation stamp from "
+ oldgs
+ " to "
+ newblock.getGenerationStamp());
}
// update length
if (newblock.getNumBytes() > oldblock.getNumBytes()) {
throw new IOException(
"Cannot update block file (="
+ blockFile
+ ") length from "
+ oldblock.getNumBytes()
+ " to "
+ newblock.getNumBytes());
}
if (newblock.getNumBytes() < oldblock.getNumBytes()) {
truncateBlock(blockFile, tmpMetaFile, oldblock.getNumBytes(), newblock.getNumBytes());
}
// rename the tmp file to the new meta file (with new generation stamp)
File newMetaFile = getMetaFile(blockFile, newblock);
if (!tmpMetaFile.renameTo(newMetaFile)) {
throw new IOException("Cannot rename tmp meta file to " + newMetaFile);
}
updateBlockMap(ongoingCreates, oldblock, newblock);
updateBlockMap(volumeMap, oldblock, newblock);
// paranoia! verify that the contents of the stored block
// matches the block file on disk.
validateBlockMetadata(newblock);
return null;
}
public void map(
WritableComparable<?> key,
Text value,
OutputCollector<Text, CrawlDatum> output,
Reporter reporter)
throws IOException {
String url = value.toString(); // value is line of text
if (url != null && url.trim().startsWith("#")) {
/* Ignore line that start with # */
return;
}
// if tabs : metadata that could be stored
// must be name=value and separated by \t
float customScore = -1f;
int customInterval = interval;
int fixedInterval = -1;
Map<String, String> metadata = new TreeMap<String, String>();
if (url.indexOf("\t") != -1) {
String[] splits = url.split("\t");
url = splits[0];
for (int s = 1; s < splits.length; s++) {
// find separation between name and value
int indexEquals = splits[s].indexOf("=");
if (indexEquals == -1) {
// skip anything without a =
continue;
}
String metaname = splits[s].substring(0, indexEquals);
String metavalue = splits[s].substring(indexEquals + 1);
if (metaname.equals(nutchScoreMDName)) {
try {
customScore = Float.parseFloat(metavalue);
} catch (NumberFormatException nfe) {
}
} else if (metaname.equals(nutchFetchIntervalMDName)) {
try {
customInterval = Integer.parseInt(metavalue);
} catch (NumberFormatException nfe) {
}
} else if (metaname.equals(nutchFixedFetchIntervalMDName)) {
try {
fixedInterval = Integer.parseInt(metavalue);
} catch (NumberFormatException nfe) {
}
} else metadata.put(metaname, metavalue);
}
}
try {
url = urlNormalizers.normalize(url, URLNormalizers.SCOPE_INJECT);
url = filters.filter(url); // filter the url
} catch (Exception e) {
if (LOG.isWarnEnabled()) {
LOG.warn("Skipping " + url + ":" + e);
}
url = null;
}
if (url == null) {
reporter.getCounter("injector", "urls_filtered").increment(1);
} else { // if it passes
value.set(url); // collect it
CrawlDatum datum = new CrawlDatum();
datum.setStatus(CrawlDatum.STATUS_INJECTED);
// Is interval custom? Then set as meta data
if (fixedInterval > -1) {
// Set writable using float. Flaot is used by AdaptiveFetchSchedule
datum
.getMetaData()
.put(Nutch.WRITABLE_FIXED_INTERVAL_KEY, new FloatWritable(fixedInterval));
datum.setFetchInterval(fixedInterval);
} else {
datum.setFetchInterval(customInterval);
}
datum.setFetchTime(curTime);
// now add the metadata
Iterator<String> keysIter = metadata.keySet().iterator();
while (keysIter.hasNext()) {
String keymd = keysIter.next();
String valuemd = metadata.get(keymd);
datum.getMetaData().put(new Text(keymd), new Text(valuemd));
}
if (customScore != -1) datum.setScore(customScore);
else datum.setScore(scoreInjected);
try {
scfilters.injectedScore(value, datum);
} catch (ScoringFilterException e) {
if (LOG.isWarnEnabled()) {
LOG.warn(
"Cannot filter injected score for url "
+ url
+ ", using default ("
+ e.getMessage()
+ ")");
}
}
reporter.getCounter("injector", "urls_injected").increment(1);
output.collect(value, datum);
}
}
/**
* A tracker wants to know if there's a Task to run. Returns a task we'd like the TaskTracker to
* execute right now.
*
* <p>Eventually this function should compute load on the various TaskTrackers, and incorporate
* knowledge of DFS file placement. But for right now, it just grabs a single item out of the
* pending task list and hands it back.
*/
public synchronized Task pollForNewTask(String taskTracker) {
//
// Compute average map and reduce task numbers across pool
//
int avgMaps = 0;
int avgReduces = 0;
int numTaskTrackers;
TaskTrackerStatus tts;
synchronized (taskTrackers) {
numTaskTrackers = taskTrackers.size();
tts = (TaskTrackerStatus) taskTrackers.get(taskTracker);
}
if (numTaskTrackers > 0) {
avgMaps = totalMaps / numTaskTrackers;
avgReduces = totalReduces / numTaskTrackers;
}
int totalCapacity = numTaskTrackers * maxCurrentTasks;
//
// Get map + reduce counts for the current tracker.
//
if (tts == null) {
LOG.warning("Unknown task tracker polling; ignoring: " + taskTracker);
return null;
}
int numMaps = tts.countMapTasks();
int numReduces = tts.countReduceTasks();
//
// In the below steps, we allocate first a map task (if appropriate),
// and then a reduce task if appropriate. We go through all jobs
// in order of job arrival; jobs only get serviced if their
// predecessors are serviced, too.
//
//
// We hand a task to the current taskTracker if the given machine
// has a workload that's equal to or less than the averageMaps
// +/- TASK_ALLOC_EPSILON. (That epsilon is in place in case
// there is an odd machine that is failing for some reason but
// has not yet been removed from the pool, making capacity seem
// larger than it really is.)
//
synchronized (jobsByArrival) {
if ((numMaps < maxCurrentTasks) && (numMaps <= (avgMaps + TASK_ALLOC_EPSILON))) {
int totalNeededMaps = 0;
for (Iterator it = jobsByArrival.iterator(); it.hasNext(); ) {
JobInProgress job = (JobInProgress) it.next();
if (job.getStatus().getRunState() != JobStatus.RUNNING) {
continue;
}
Task t = job.obtainNewMapTask(taskTracker, tts);
if (t != null) {
return t;
}
//
// Beyond the highest-priority task, reserve a little
// room for failures and speculative executions; don't
// schedule tasks to the hilt.
//
totalNeededMaps += job.desiredMaps();
double padding = 0;
if (totalCapacity > MIN_SLOTS_FOR_PADDING) {
padding = Math.min(maxCurrentTasks, totalNeededMaps * PAD_FRACTION);
}
if (totalNeededMaps + padding >= totalCapacity) {
break;
}
}
}
//
// Same thing, but for reduce tasks
//
if ((numReduces < maxCurrentTasks) && (numReduces <= (avgReduces + TASK_ALLOC_EPSILON))) {
int totalNeededReduces = 0;
for (Iterator it = jobsByArrival.iterator(); it.hasNext(); ) {
JobInProgress job = (JobInProgress) it.next();
if (job.getStatus().getRunState() != JobStatus.RUNNING) {
continue;
}
Task t = job.obtainNewReduceTask(taskTracker, tts);
if (t != null) {
return t;
}
//
// Beyond the highest-priority task, reserve a little
// room for failures and speculative executions; don't
// schedule tasks to the hilt.
//
totalNeededReduces += job.desiredReduces();
double padding = 0;
if (totalCapacity > MIN_SLOTS_FOR_PADDING) {
padding = Math.min(maxCurrentTasks, totalNeededReduces * PAD_FRACTION);
}
if (totalNeededReduces + padding >= totalCapacity) {
break;
}
}
}
}
return null;
}