public void inject(Path crawlDb, Path urlDir) throws IOException {
SimpleDateFormat sdf = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
long start = System.currentTimeMillis();
if (LOG.isInfoEnabled()) {
LOG.info("Injector: starting at " + sdf.format(start));
LOG.info("Injector: crawlDb: " + crawlDb);
LOG.info("Injector: urlDir: " + urlDir);
}
Path tempDir =
new Path(
getConf().get("mapred.temp.dir", ".")
+ "/inject-temp-"
+ Integer.toString(new Random().nextInt(Integer.MAX_VALUE)));
// map text input file to a <url,CrawlDatum> file
if (LOG.isInfoEnabled()) {
LOG.info("Injector: Converting injected urls to crawl db entries.");
}
JobConf sortJob = new NutchJob(getConf());
sortJob.setJobName("inject " + urlDir);
FileInputFormat.addInputPath(sortJob, urlDir);
sortJob.setMapperClass(InjectMapper.class);
FileOutputFormat.setOutputPath(sortJob, tempDir);
sortJob.setOutputFormat(SequenceFileOutputFormat.class);
sortJob.setOutputKeyClass(Text.class);
sortJob.setOutputValueClass(CrawlDatum.class);
sortJob.setLong("injector.current.time", System.currentTimeMillis());
RunningJob mapJob = JobClient.runJob(sortJob);
long urlsInjected = mapJob.getCounters().findCounter("injector", "urls_injected").getValue();
long urlsFiltered = mapJob.getCounters().findCounter("injector", "urls_filtered").getValue();
LOG.info("Injector: total number of urls rejected by filters: " + urlsFiltered);
LOG.info(
"Injector: total number of urls injected after normalization and filtering: "
+ urlsInjected);
// merge with existing crawl db
if (LOG.isInfoEnabled()) {
LOG.info("Injector: Merging injected urls into crawl db.");
}
JobConf mergeJob = CrawlDb.createJob(getConf(), crawlDb);
FileInputFormat.addInputPath(mergeJob, tempDir);
mergeJob.setReducerClass(InjectReducer.class);
JobClient.runJob(mergeJob);
CrawlDb.install(mergeJob, crawlDb);
// clean up
FileSystem fs = FileSystem.get(getConf());
fs.delete(tempDir, true);
long end = System.currentTimeMillis();
LOG.info(
"Injector: finished at "
+ sdf.format(end)
+ ", elapsed: "
+ TimingUtil.elapsedTime(start, end));
}
public void configure(JobConf job) {
this.jobConf = job;
urlNormalizers = new URLNormalizers(job, URLNormalizers.SCOPE_INJECT);
interval = jobConf.getInt("db.fetch.interval.default", 2592000);
filters = new URLFilters(jobConf);
scfilters = new ScoringFilters(jobConf);
scoreInjected = jobConf.getFloat("db.score.injected", 1.0f);
curTime = job.getLong("injector.current.time", System.currentTimeMillis());
}
public void bumpProgress() {
numWritten++;
if (numWritten % 25000 == 0) {
long now = System.currentTimeMillis();
long delta = now - lastCheckpoint;
lastCheckpoint = now;
LOG.info("Wrote last 25000 records in " + delta + " ms");
localManager.progress();
}
}
private static void analyzeResult(FileSystem fs, int testType, long execTime, String resFileName)
throws IOException {
Path reduceFile;
if (testType == TEST_TYPE_WRITE) reduceFile = new Path(WRITE_DIR, "part-00000");
else reduceFile = new Path(READ_DIR, "part-00000");
DataInputStream in;
in = new DataInputStream(fs.open(reduceFile));
BufferedReader lines;
lines = new BufferedReader(new InputStreamReader(in));
long tasks = 0;
long size = 0;
long time = 0;
float rate = 0;
float sqrate = 0;
String line;
while ((line = lines.readLine()) != null) {
StringTokenizer tokens = new StringTokenizer(line, " \t\n\r\f%");
String attr = tokens.nextToken();
if (attr.endsWith(":tasks")) tasks = Long.parseLong(tokens.nextToken());
else if (attr.endsWith(":size")) size = Long.parseLong(tokens.nextToken());
else if (attr.endsWith(":time")) time = Long.parseLong(tokens.nextToken());
else if (attr.endsWith(":rate")) rate = Float.parseFloat(tokens.nextToken());
else if (attr.endsWith(":sqrate")) sqrate = Float.parseFloat(tokens.nextToken());
}
double med = rate / 1000 / tasks;
double stdDev = Math.sqrt(Math.abs(sqrate / 1000 / tasks - med * med));
String resultLines[] = {
"----- DFSCIOTest ----- : "
+ ((testType == TEST_TYPE_WRITE)
? "write"
: (testType == TEST_TYPE_READ) ? "read" : "unknown"),
" Date & time: " + new Date(System.currentTimeMillis()),
" Number of files: " + tasks,
"Total MBytes processed: " + size / MEGA,
" Throughput mb/sec: " + size * 1000.0 / (time * MEGA),
"Average IO rate mb/sec: " + med,
" Std IO rate deviation: " + stdDev,
" Test exec time sec: " + (float) execTime / 1000,
""
};
PrintStream res = new PrintStream(new FileOutputStream(new File(resFileName), true));
for (int i = 0; i < resultLines.length; i++) {
LOG.info(resultLines[i]);
res.println(resultLines[i]);
}
}
public int run(String[] args) throws Exception {
if (args.length < 4) {
System.out.println("ERROR: Please Enter args : input output type(text|seq) splitChar(9=\t)");
return JobClient.SUCCESS;
}
String input = args[0];
String output = args[1];
String type = args[2];
String splitChar = args[3];
JobConf config = new JobConf(getConf(), getClass());
config.set("user.split", splitChar);
config.setJobName("File Filter -" + System.currentTimeMillis());
config.setNumReduceTasks(10);
config.setReducerClass(IdentityReducer.class);
config.setMapperClass(FileTestMapper.class);
if ("text".equals(type)) {
config.setInputFormat(TextInputFormat.class);
TextInputFormat.addInputPath(config, new Path(input));
} else {
config.setInputFormat(SequenceFileInputFormat.class);
SequenceFileInputFormat.addInputPath(config, new Path(input));
}
config.setMapOutputKeyClass(Text.class);
config.setMapOutputValueClass(Text.class);
config.setOutputKeyClass(Text.class);
config.setOutputValueClass(Text.class);
// if output path exists then return
FileSystem fs = FileSystem.get(config);
Path outputPath = new Path(output);
FileOutputFormat.setOutputPath(config, outputPath);
if (!fs.exists(outputPath)) {
JobClient.runJob(config);
} else {
System.out.println("You has finished this job today ! " + outputPath);
}
return JobClient.SUCCESS;
}
public int run(String[] args) throws Exception {
if (args.length < 1) {
args = new String[] {DateStringUtils.now()};
System.out.println(
"ERROR: Please Enter Date , eg. 20101010 ! now use default => " + DateStringUtils.now());
}
JobConf config = new JobConf(getConf(), getClass());
config.set("user.args", Utils.asString(args));
config.setJobName(getClass() + "-" + System.currentTimeMillis());
config.setNumReduceTasks(100);
config.setMapperClass(getClass());
config.setReducerClass(getClass());
config.setInputFormat(getInputFormat());
config.setMapOutputKeyClass(Text.class);
config.setMapOutputValueClass(Text.class);
// add input paths
for (String path : getInputPath(args)) {
if (TextInputFormat.class.equals(getInputFormat())) {
TextInputFormat.addInputPath(config, new Path(path));
} else if (SequenceFileInputFormat.class.equals(getInputFormat())) {
SequenceFileInputFormat.addInputPath(config, new Path(path));
}
}
config.setOutputKeyClass(Text.class);
config.setOutputValueClass(Text.class);
// if output path exists then return
FileSystem fs = FileSystem.get(config);
Path outputPath = new Path(getOutputPath(args));
FileOutputFormat.setOutputPath(config, outputPath);
if (!fs.exists(outputPath)) {
JobClient.runJob(config);
} else {
System.out.println("You has finished this job today ! " + outputPath);
}
return JobClient.SUCCESS;
}
public void testFsCache() throws Exception {
{
long now = System.currentTimeMillis();
String[] users = new String[] {"foo", "bar"};
final Configuration conf = new Configuration();
FileSystem[] fs = new FileSystem[users.length];
for (int i = 0; i < users.length; i++) {
UserGroupInformation ugi = UserGroupInformation.createRemoteUser(users[i]);
fs[i] =
ugi.doAs(
new PrivilegedExceptionAction<FileSystem>() {
public FileSystem run() throws IOException {
return FileSystem.get(conf);
}
});
for (int j = 0; j < i; j++) {
assertFalse(fs[j] == fs[i]);
}
}
FileSystem.closeAll();
}
{
try {
runTestCache(HdfsClientConfigKeys.DFS_NAMENODE_RPC_PORT_DEFAULT);
} catch (java.net.BindException be) {
LOG.warn(
"Cannot test HdfsClientConfigKeys.DFS_NAMENODE_RPC_PORT_DEFAULT (="
+ HdfsClientConfigKeys.DFS_NAMENODE_RPC_PORT_DEFAULT
+ ")",
be);
}
runTestCache(0);
}
}
public static void main(String[] args) {
int testType = TEST_TYPE_READ;
int bufferSize = DEFAULT_BUFFER_SIZE;
int fileSize = 1;
int nrFiles = 1;
String resFileName = DEFAULT_RES_FILE_NAME;
boolean isSequential = false;
String version = "DFSCIOTest.0.0.1";
String usage =
"Usage: DFSCIOTest -read | -write | -clean [-nrFiles N] [-fileSize MB] [-resFile resultFileName] [-bufferSize Bytes] ";
System.out.println(version);
if (args.length == 0) {
System.err.println(usage);
System.exit(-1);
}
for (int i = 0; i < args.length; i++) { // parse command line
if (args[i].startsWith("-r")) {
testType = TEST_TYPE_READ;
} else if (args[i].startsWith("-w")) {
testType = TEST_TYPE_WRITE;
} else if (args[i].startsWith("-clean")) {
testType = TEST_TYPE_CLEANUP;
} else if (args[i].startsWith("-seq")) {
isSequential = true;
} else if (args[i].equals("-nrFiles")) {
nrFiles = Integer.parseInt(args[++i]);
} else if (args[i].equals("-fileSize")) {
fileSize = Integer.parseInt(args[++i]);
} else if (args[i].equals("-bufferSize")) {
bufferSize = Integer.parseInt(args[++i]);
} else if (args[i].equals("-resFile")) {
resFileName = args[++i];
}
}
LOG.info("nrFiles = " + nrFiles);
LOG.info("fileSize (MB) = " + fileSize);
LOG.info("bufferSize = " + bufferSize);
try {
fsConfig.setInt("test.io.file.buffer.size", bufferSize);
FileSystem fs = FileSystem.get(fsConfig);
if (testType != TEST_TYPE_CLEANUP) {
fs.delete(HDFS_TEST_DIR, true);
if (!fs.mkdirs(HDFS_TEST_DIR)) {
throw new IOException("Mkdirs failed to create " + HDFS_TEST_DIR.toString());
}
// Copy the executables over to the remote filesystem
String hadoopHome = System.getenv("HADOOP_PREFIX");
fs.copyFromLocalFile(
new Path(hadoopHome + "/libhdfs/libhdfs.so." + HDFS_LIB_VERSION), HDFS_SHLIB);
fs.copyFromLocalFile(new Path(hadoopHome + "/libhdfs/hdfs_read"), HDFS_READ);
fs.copyFromLocalFile(new Path(hadoopHome + "/libhdfs/hdfs_write"), HDFS_WRITE);
}
if (isSequential) {
long tStart = System.currentTimeMillis();
sequentialTest(fs, testType, fileSize, nrFiles);
long execTime = System.currentTimeMillis() - tStart;
String resultLine = "Seq Test exec time sec: " + (float) execTime / 1000;
LOG.info(resultLine);
return;
}
if (testType == TEST_TYPE_CLEANUP) {
cleanup(fs);
return;
}
createControlFile(fs, fileSize, nrFiles);
long tStart = System.currentTimeMillis();
if (testType == TEST_TYPE_WRITE) writeTest(fs);
if (testType == TEST_TYPE_READ) readTest(fs);
long execTime = System.currentTimeMillis() - tStart;
analyzeResult(fs, testType, execTime, resFileName);
} catch (Exception e) {
System.err.print(e.getLocalizedMessage());
System.exit(-1);
}
}
public class ElephantRecordWriter
implements RecordWriter<IntWritable, ElephantRecordWritable>, Closeable {
FileSystem fileSystem;
Args args;
Map<Integer, Persistence> lps = new HashMap<Integer, Persistence>();
Progressable progressable;
LocalElephantManager localManager;
int numWritten = 0;
long lastCheckpoint = System.currentTimeMillis();
public ElephantRecordWriter(Configuration conf, Args args, Progressable progressable)
throws IOException {
fileSystem = Utils.getFS(args.outputDirHdfs, conf);
this.args = args;
this.progressable = progressable;
localManager =
new LocalElephantManager(fileSystem, args.spec, LocalElephantManager.getTmpDirs(conf));
}
private Persistence retrieveShard(int shardIdx) throws IOException {
Persistence lp = null;
if (lps.containsKey(shardIdx)) {
lp = lps.get(shardIdx);
} else {
String localShard = localManager.downloadRemoteShard("" + shardIdx, null);
Coordinator fact = args.spec.getCoordinator();
lp = fact.openPersistenceForAppend(localShard, args.spec.getPersistenceOptions());
lps.put(shardIdx, lp);
progress();
}
return lp;
}
public void write(IntWritable shard, ElephantRecordWritable carrier) throws IOException {
Persistence lp = retrieveShard(shard.get());
NewKeyValDocument doc = new NewKeyValDocument(carrier.key, carrier.value);
lp.index(doc);
bumpProgress();
}
public void bumpProgress() {
numWritten++;
if (numWritten % 25000 == 0) {
long now = System.currentTimeMillis();
long delta = now - lastCheckpoint;
lastCheckpoint = now;
LOG.info("Wrote last 25000 records in " + delta + " ms");
localManager.progress();
}
}
public void close() throws IOException {
close(null);
}
public void close(Reporter reporter) throws IOException {
for (Integer shard : lps.keySet()) {
String lpDir = localManager.localTmpDir("" + shard);
LOG.info("Closing LP for shard " + shard + " at " + lpDir);
lps.get(shard).close();
LOG.info("Closed LP for shard " + shard + " at " + lpDir);
progress();
String remoteDir = args.outputDirHdfs + "/" + shard;
// Do all this stuff to ensure that S3 actually does delete
int deleteAttempt = 4;
while (fileSystem.exists(new Path(remoteDir)) && deleteAttempt > 0) {
LOG.info("Deleting existing shard " + shard + " at " + remoteDir);
fileSystem.delete(new Path(remoteDir), true);
--deleteAttempt;
}
if (fileSystem.exists(new Path(remoteDir)) && deleteAttempt == 0) {
throw new IOException(
"Failed to delete shard "
+ shard
+ " at "
+ remoteDir
+ " after "
+ deleteAttempt
+ " attempts!");
} else {
LOG.info("Deleted existing shard " + shard + " at " + remoteDir);
}
LOG.info("Copying " + lpDir + " to " + remoteDir);
fileSystem.copyFromLocalFile(new Path(lpDir), new Path(remoteDir));
LOG.info("Copied " + lpDir + " to " + remoteDir);
progress();
}
localManager.cleanup();
}
private void progress() {
if (progressable != null) progressable.progress();
}
}
