public static void main(String[] args) throws IOException {
if (args.length != 3) {
System.out.println("Parameters: inputDir outputDir parallel");
System.exit(1);
}
String inputDir = args[0];
String outputDir = args[1];
String parallel = args[2];
JobConf lp = new JobConf(L10.class);
lp.setJobName("L10 Load Page Views");
lp.setInputFormat(TextInputFormat.class);
lp.setOutputKeyClass(MyType.class);
lp.setOutputValueClass(Text.class);
lp.setMapperClass(ReadPageViews.class);
lp.setReducerClass(Group.class);
lp.setPartitionerClass(MyPartitioner.class);
Properties props = System.getProperties();
for (Map.Entry<Object, Object> entry : props.entrySet()) {
lp.set((String) entry.getKey(), (String) entry.getValue());
}
FileInputFormat.addInputPath(lp, new Path(inputDir + "/page_views"));
FileOutputFormat.setOutputPath(lp, new Path(outputDir + "/L10out"));
// Hardcode the parallel to 40 since MyPartitioner assumes it
lp.setNumReduceTasks(40);
Job group = new Job(lp);
JobControl jc = new JobControl("L10 join");
jc.addJob(group);
new Thread(jc).start();
int i = 0;
while (!jc.allFinished()) {
ArrayList<Job> failures = jc.getFailedJobs();
if (failures != null && failures.size() > 0) {
for (Job failure : failures) {
System.err.println(failure.getMessage());
}
break;
}
try {
Thread.sleep(5000);
} catch (InterruptedException e) {
}
if (i % 10000 == 0) {
System.out.println("Running jobs");
ArrayList<Job> running = jc.getRunningJobs();
if (running != null && running.size() > 0) {
for (Job r : running) {
System.out.println(r.getJobName());
}
}
System.out.println("Ready jobs");
ArrayList<Job> ready = jc.getReadyJobs();
if (ready != null && ready.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
System.out.println("Waiting jobs");
ArrayList<Job> waiting = jc.getWaitingJobs();
if (waiting != null && waiting.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
System.out.println("Successful jobs");
ArrayList<Job> success = jc.getSuccessfulJobs();
if (success != null && success.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
}
i++;
}
ArrayList<Job> failures = jc.getFailedJobs();
if (failures != null && failures.size() > 0) {
for (Job failure : failures) {
System.err.println(failure.getMessage());
}
}
jc.stop();
}
public static void main(String[] args) throws IOException {
JobConf lp = new JobConf(L4.class);
lp.setJobName("Load Page Views");
lp.setInputFormat(TextInputFormat.class);
lp.setOutputKeyClass(Text.class);
lp.setOutputValueClass(Text.class);
lp.setMapperClass(ReadPageViews.class);
lp.setCombinerClass(Combiner.class);
lp.setReducerClass(Group.class);
Properties props = System.getProperties();
String dataDir = props.getProperty("PIGMIX_DIR", "/user/pig/tests/data/pigmix");
for (Map.Entry<Object, Object> entry : props.entrySet()) {
lp.set((String) entry.getKey(), (String) entry.getValue());
}
FileInputFormat.addInputPath(lp, new Path(dataDir, "page_views"));
FileOutputFormat.setOutputPath(
lp, new Path("/user/" + System.getProperty("user.name") + "/L4out"));
lp.setNumReduceTasks(40);
Job group = new Job(lp);
JobControl jc = new JobControl("L4 join");
jc.addJob(group);
new Thread(jc).start();
int i = 0;
while (!jc.allFinished()) {
ArrayList<Job> failures = jc.getFailedJobs();
if (failures != null && failures.size() > 0) {
for (Job failure : failures) {
System.err.println(failure.getMessage());
}
break;
}
try {
Thread.sleep(5000);
} catch (InterruptedException e) {
}
if (i % 10000 == 0) {
System.out.println("Running jobs");
ArrayList<Job> running = jc.getRunningJobs();
if (running != null && running.size() > 0) {
for (Job r : running) {
System.out.println(r.getJobName());
}
}
System.out.println("Ready jobs");
ArrayList<Job> ready = jc.getReadyJobs();
if (ready != null && ready.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
System.out.println("Waiting jobs");
ArrayList<Job> waiting = jc.getWaitingJobs();
if (waiting != null && waiting.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
System.out.println("Successful jobs");
ArrayList<Job> success = jc.getSuccessfulJobs();
if (success != null && success.size() > 0) {
for (Job r : ready) {
System.out.println(r.getJobName());
}
}
}
i++;
}
ArrayList<Job> failures = jc.getFailedJobs();
if (failures != null && failures.size() > 0) {
for (Job failure : failures) {
System.err.println(failure.getMessage());
}
}
jc.stop();
}