blob: 469705443458f9652a1c85428f4e9d0570afec8c [file] [log] [blame]
/**
* 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.hdfs;
import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.io.InputStream;
import java.net.InetSocketAddress;
import java.net.URL;
import java.net.URLConnection;
import java.security.PrivilegedExceptionAction;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Random;
import java.util.concurrent.TimeoutException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.BlockLocation;
import org.apache.hadoop.fs.CommonConfigurationKeys;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.FileSystem.Statistics;
import org.apache.hadoop.hdfs.DFSClient.DFSDataInputStream;
import org.apache.hadoop.hdfs.protocol.ClientDatanodeProtocol;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.protocol.DatanodeID;
import org.apache.hadoop.hdfs.protocol.LocatedBlock;
import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
import org.apache.hadoop.hdfs.security.token.block.BlockTokenIdentifier;
import org.apache.hadoop.hdfs.server.namenode.DatanodeDescriptor;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.security.ShellBasedUnixGroupsMapping;
import org.apache.hadoop.security.token.Token;
import org.apache.hadoop.security.UserGroupInformation;
/** Utilities for HDFS tests */
public class DFSTestUtil {
private static Random gen = new Random();
private static String[] dirNames = {
"zero", "one", "two", "three", "four", "five", "six", "seven", "eight", "nine"
};
private int maxLevels;// = 3;
private int maxSize;// = 8*1024;
private int nFiles;
private MyFile[] files;
/** Creates a new instance of DFSTestUtil
*
* @param testName Name of the test from where this utility is used
* @param nFiles Number of files to be created
* @param maxLevels Maximum number of directory levels
* @param maxSize Maximum size for file
*/
public DFSTestUtil(String testName, int nFiles, int maxLevels, int maxSize) {
this.nFiles = nFiles;
this.maxLevels = maxLevels;
this.maxSize = maxSize;
}
/** class MyFile contains enough information to recreate the contents of
* a single file.
*/
private class MyFile {
private String name = "";
private int size;
private long seed;
MyFile() {
int nLevels = gen.nextInt(maxLevels);
if (nLevels != 0) {
int[] levels = new int[nLevels];
for (int idx = 0; idx < nLevels; idx++) {
levels[idx] = gen.nextInt(10);
}
StringBuffer sb = new StringBuffer();
for (int idx = 0; idx < nLevels; idx++) {
sb.append(dirNames[levels[idx]]);
sb.append("/");
}
name = sb.toString();
}
long fidx = -1;
while (fidx < 0) { fidx = gen.nextLong(); }
name = name + Long.toString(fidx);
size = gen.nextInt(maxSize);
seed = gen.nextLong();
}
String getName() { return name; }
int getSize() { return size; }
long getSeed() { return seed; }
}
public void createFiles(FileSystem fs, String topdir) throws IOException {
createFiles(fs, topdir, (short)3);
}
/** create nFiles with random names and directory hierarchies
* with random (but reproducible) data in them.
*/
public void createFiles(FileSystem fs, String topdir,
short replicationFactor) throws IOException {
files = new MyFile[nFiles];
for (int idx = 0; idx < nFiles; idx++) {
files[idx] = new MyFile();
}
Path root = new Path(topdir);
for (int idx = 0; idx < nFiles; idx++) {
createFile(fs, new Path(root, files[idx].getName()), files[idx].getSize(),
replicationFactor, files[idx].getSeed());
}
}
public static String readFile(FileSystem fs, Path fileName) throws IOException {
ByteArrayOutputStream os = new ByteArrayOutputStream();
IOUtils.copyBytes(fs.open(fileName), os, 1024, true);
return os.toString();
}
public static void createFile(FileSystem fs, Path fileName, long fileLen,
short replFactor, long seed) throws IOException {
if (!fs.mkdirs(fileName.getParent())) {
throw new IOException("Mkdirs failed to create " +
fileName.getParent().toString());
}
FSDataOutputStream out = null;
try {
out = fs.create(fileName, replFactor);
byte[] toWrite = new byte[1024];
Random rb = new Random(seed);
long bytesToWrite = fileLen;
while (bytesToWrite>0) {
rb.nextBytes(toWrite);
int bytesToWriteNext = (1024<bytesToWrite)?1024:(int)bytesToWrite;
out.write(toWrite, 0, bytesToWriteNext);
bytesToWrite -= bytesToWriteNext;
}
out.close();
out = null;
} finally {
IOUtils.closeStream(out);
}
}
/** check if the files have been copied correctly. */
public boolean checkFiles(FileSystem fs, String topdir) throws IOException {
Path root = new Path(topdir);
for (int idx = 0; idx < nFiles; idx++) {
Path fPath = new Path(root, files[idx].getName());
FSDataInputStream in = fs.open(fPath);
byte[] toRead = new byte[files[idx].getSize()];
byte[] toCompare = new byte[files[idx].getSize()];
Random rb = new Random(files[idx].getSeed());
rb.nextBytes(toCompare);
in.readFully(0, toRead);
in.close();
for (int i = 0; i < toRead.length; i++) {
if (toRead[i] != toCompare[i]) {
return false;
}
}
toRead = null;
toCompare = null;
}
return true;
}
void setReplication(FileSystem fs, String topdir, short value)
throws IOException {
Path root = new Path(topdir);
for (int idx = 0; idx < nFiles; idx++) {
Path fPath = new Path(root, files[idx].getName());
fs.setReplication(fPath, value);
}
}
/*
* Waits for the replication factor of all files to reach the
* specified target.
*/
public void waitReplication(FileSystem fs, String topdir, short value)
throws IOException {
Path root = new Path(topdir);
/** wait for the replication factor to settle down */
for (int idx = 0; idx < nFiles; idx++) {
waitReplication(fs, new Path(root, files[idx].getName()), value);
}
}
/*
* Check if the given block in the given file is corrupt.
*/
public static boolean allBlockReplicasCorrupt(MiniDFSCluster cluster,
Path file, int blockNo) throws IOException {
DFSClient client = new DFSClient(new InetSocketAddress("localhost",
cluster.getNameNodePort()), cluster.getConfiguration());
LocatedBlocks blocks;
try {
blocks = client.getNamenode().getBlockLocations(
file.toString(), 0, Long.MAX_VALUE);
} finally {
client.close();
}
return blocks.get(blockNo).isCorrupt();
}
/*
* Wait up to 20s for the given block to be replicated across
* the requested number of racks, with the requested number of
* replicas, and the requested number of replicas still needed.
*/
public static void waitForReplication(MiniDFSCluster cluster, Block b,
int racks, int replicas, int neededReplicas)
throws IOException, TimeoutException, InterruptedException {
int curRacks = 0;
int curReplicas = 0;
int curNeededReplicas = 0;
int count = 0;
final int ATTEMPTS = 20;
do {
Thread.sleep(1000);
int []r = NameNodeAdapter.getReplicaInfo(cluster.getNameNode(), b);
curRacks = r[0];
curReplicas = r[1];
curNeededReplicas = r[2];
count++;
} while ((curRacks != racks ||
curReplicas != replicas ||
curNeededReplicas != neededReplicas) && count < ATTEMPTS);
if (count == ATTEMPTS) {
throw new TimeoutException("Timed out waiting for replication."
+ " Needed replicas = "+neededReplicas
+ " Cur needed replicas = "+curNeededReplicas
+ " Replicas = "+replicas+" Cur replicas = "+curReplicas
+ " Racks = "+racks+" Cur racks = "+curRacks);
}
}
/*
* Keep accessing the given file until the namenode reports that the
* given block in the file contains the given number of corrupt replicas.
*/
public static void waitCorruptReplicas(FileSystem fs, FSNamesystem ns,
Path file, Block b, int corruptRepls)
throws IOException, TimeoutException {
int count = 0;
final int ATTEMPTS = 50;
int repls = ns.numCorruptReplicas(b);
while (repls != corruptRepls && count < ATTEMPTS) {
try {
IOUtils.copyBytes(fs.open(file), new IOUtils.NullOutputStream(),
512, true);
} catch (IOException e) {
// Swallow exceptions
}
System.out.println("Waiting for "+corruptRepls+" corrupt replicas");
repls = ns.numCorruptReplicas(b);
count++;
}
if (count == ATTEMPTS) {
throw new TimeoutException("Timed out waiting for corrupt replicas."
+ " Waiting for "+corruptRepls+", but only found "+repls);
}
}
/*
* Wait up to 20s for the given DN (host:port) to be decommissioned.
*/
public static void waitForDecommission(FileSystem fs, String name)
throws IOException, InterruptedException, TimeoutException {
DatanodeInfo dn = null;
int count = 0;
final int ATTEMPTS = 20;
do {
Thread.sleep(1000);
DistributedFileSystem dfs = (DistributedFileSystem)fs;
for (DatanodeInfo info : dfs.getDataNodeStats()) {
if (name.equals(info.getName())) {
dn = info;
}
}
count++;
} while ((dn == null ||
dn.isDecommissionInProgress() ||
!dn.isDecommissioned()) && count < ATTEMPTS);
if (count == ATTEMPTS) {
throw new TimeoutException("Timed out waiting for datanode "
+ name + " to decommission.");
}
}
/*
* Returns the index of the first datanode which has a copy
* of the given block, or -1 if no such datanode exists.
*/
public static int firstDnWithBlock(MiniDFSCluster cluster, Block b)
throws IOException {
int numDatanodes = cluster.getDataNodes().size();
for (int i = 0; i < numDatanodes; i++) {
String blockContent = cluster.readBlockOnDataNode(i, b.getBlockName());
if (blockContent != null) {
return i;
}
}
return -1;
}
/*
* Return the total capacity of all live DNs.
*/
public static long getLiveDatanodeCapacity(FSNamesystem ns) {
ArrayList<DatanodeDescriptor> live = new ArrayList<DatanodeDescriptor>();
ArrayList<DatanodeDescriptor> dead = new ArrayList<DatanodeDescriptor>();
ns.DFSNodesStatus(live, dead);
long capacity = 0;
for (final DatanodeDescriptor dn : live) {
capacity += dn.getCapacity();
}
return capacity;
}
/*
* Return the capacity of the given live DN.
*/
public static long getDatanodeCapacity(FSNamesystem ns, int index) {
ArrayList<DatanodeDescriptor> live = new ArrayList<DatanodeDescriptor>();
ArrayList<DatanodeDescriptor> dead = new ArrayList<DatanodeDescriptor>();
ns.DFSNodesStatus(live, dead);
return live.get(index).getCapacity();
}
/*
* Wait for the given # live/dead DNs, total capacity, and # vol failures.
*/
public static void waitForDatanodeStatus(FSNamesystem ns, int expectedLive,
int expectedDead, long expectedVolFails, long expectedTotalCapacity,
long timeout) throws InterruptedException, TimeoutException {
ArrayList<DatanodeDescriptor> live = new ArrayList<DatanodeDescriptor>();
ArrayList<DatanodeDescriptor> dead = new ArrayList<DatanodeDescriptor>();
final int ATTEMPTS = 10;
int count = 0;
long currTotalCapacity = 0;
int volFails = 0;
do {
Thread.sleep(timeout);
live.clear();
dead.clear();
ns.DFSNodesStatus(live, dead);
currTotalCapacity = 0;
volFails = 0;
for (final DatanodeDescriptor dd : live) {
currTotalCapacity += dd.getCapacity();
volFails += dd.getVolumeFailures();
}
count++;
} while ((expectedLive != live.size() ||
expectedDead != dead.size() ||
expectedTotalCapacity != currTotalCapacity ||
expectedVolFails != volFails)
&& count < ATTEMPTS);
if (count == ATTEMPTS) {
throw new TimeoutException("Timed out waiting for capacity."
+ " Live = "+live.size()+" Expected = "+expectedLive
+ " Dead = "+dead.size()+" Expected = "+expectedDead
+ " Total capacity = "+currTotalCapacity
+ " Expected = "+expectedTotalCapacity
+ " Vol Fails = "+volFails+" Expected = "+expectedVolFails);
}
}
/** return list of filenames created as part of createFiles */
public String[] getFileNames(String topDir) {
if (nFiles == 0)
return new String[]{};
else {
String[] fileNames = new String[nFiles];
for (int idx=0; idx < nFiles; idx++) {
fileNames[idx] = topDir + "/" + files[idx].getName();
}
return fileNames;
}
}
/** wait for the file's replication to be done */
public static void waitReplication(FileSystem fs, Path fileName,
short replFactor) throws IOException {
boolean good;
do {
good = true;
BlockLocation locs[] = fs.getFileBlockLocations(
fs.getFileStatus(fileName), 0, Long.MAX_VALUE);
for (int j = 0; j < locs.length; j++) {
String[] loc = locs[j].getHosts();
if (loc.length != replFactor) {
System.out.println("File " + fileName + " has replication factor " +
loc.length);
good = false;
try {
System.out.println("Waiting for replication factor to drain");
Thread.sleep(100);
} catch (InterruptedException e) {}
break;
}
}
} while(!good);
}
/** delete directory and everything underneath it.*/
public void cleanup(FileSystem fs, String topdir) throws IOException {
Path root = new Path(topdir);
fs.delete(root, true);
files = null;
}
public static Block getFirstBlock(FileSystem fs, Path path) throws IOException {
DFSDataInputStream in =
(DFSDataInputStream) ((DistributedFileSystem)fs).open(path);
in.readByte();
return in.getCurrentBlock();
}
public static List<LocatedBlock> getAllBlocks(FSDataInputStream in)
throws IOException {
return ((DFSClient.DFSDataInputStream) in).getAllBlocks();
}
public static Token<BlockTokenIdentifier> getBlockToken(
FSDataOutputStream out) {
return ((DFSOutputStream) out.getWrappedStream()).getBlockToken();
}
public static ClientDatanodeProtocol createClientDatanodeProtocolProxy(
DatanodeID datanodeid, Configuration conf, int socketTimeout,
LocatedBlock locatedBlock)
throws IOException {
return DFSClient.createClientDatanodeProtocolProxy(
datanodeid, conf, socketTimeout, locatedBlock);
}
static void setLogLevel2All(org.apache.commons.logging.Log log) {
((org.apache.commons.logging.impl.Log4JLogger)log
).getLogger().setLevel(org.apache.log4j.Level.ALL);
}
public static String readFile(File f) throws IOException {
StringBuilder b = new StringBuilder();
BufferedReader in = new BufferedReader(new FileReader(f));
for(int c; (c = in.read()) != -1; b.append((char)c));
in.close();
return b.toString();
}
/* Write the given string to the given file */
public static void writeFile(FileSystem fs, Path p, String s)
throws IOException {
if (fs.exists(p)) {
fs.delete(p, true);
}
InputStream is = new ByteArrayInputStream(s.getBytes());
FSDataOutputStream os = fs.create(p);
IOUtils.copyBytes(is, os, s.length(), true);
}
/* Append the given string to the given file */
public static void appendFile(FileSystem fs, Path p, String s)
throws IOException {
assert fs.exists(p);
InputStream is = new ByteArrayInputStream(s.getBytes());
FSDataOutputStream os = fs.append(p);
IOUtils.copyBytes(is, os, s.length(), true);
}
// Returns url content as string.
public static String urlGet(URL url) throws IOException {
URLConnection conn = url.openConnection();
ByteArrayOutputStream out = new ByteArrayOutputStream();
IOUtils.copyBytes(conn.getInputStream(), out, 4096, true);
return out.toString();
}
/**
* mock class to get group mapping for fake users
*
*/
static class MockUnixGroupsMapping extends ShellBasedUnixGroupsMapping {
static Map<String, String []> fakeUser2GroupsMap;
private static final List<String> defaultGroups;
static {
defaultGroups = new ArrayList<String>(1);
defaultGroups.add("supergroup");
fakeUser2GroupsMap = new HashMap<String, String[]>();
}
@Override
public List<String> getGroups(String user) throws IOException {
boolean found = false;
// check to see if this is one of fake users
List<String> l = new ArrayList<String>();
for(String u : fakeUser2GroupsMap.keySet()) {
if(user.equals(u)) {
found = true;
for(String gr : fakeUser2GroupsMap.get(u)) {
l.add(gr);
}
}
}
// default
if(!found) {
l = super.getGroups(user);
if(l.size() == 0) {
System.out.println("failed to get real group for " + user +
"; using default");
return defaultGroups;
}
}
return l;
}
}
/**
* update the configuration with fake class for mapping user to groups
* @param conf
* @param map - user to groups mapping
*/
static public void updateConfWithFakeGroupMapping
(Configuration conf, Map<String, String []> map) {
if(map!=null) {
MockUnixGroupsMapping.fakeUser2GroupsMap = map;
}
// fake mapping user to groups
conf.setClass(CommonConfigurationKeys.HADOOP_SECURITY_GROUP_MAPPING,
DFSTestUtil.MockUnixGroupsMapping.class,
ShellBasedUnixGroupsMapping.class);
}
/**
* Get a FileSystem instance as specified user in a doAs block.
*/
static public FileSystem getFileSystemAs(UserGroupInformation ugi,
final Configuration conf) throws IOException,
InterruptedException {
return ugi.doAs(new PrivilegedExceptionAction<FileSystem>() {
@Override
public FileSystem run() throws Exception {
return FileSystem.get(conf);
}
});
}
public static byte[] generateSequentialBytes(int start, int length) {
byte[] result = new byte[length];
for (int i = 0; i < length; i++) {
result[i] = (byte) ((start + i) % 127);
}
return result;
}
public static Statistics getStatistics(FileSystem fs) {
return FileSystem.getStatistics(fs.getUri().getScheme(), fs.getClass());
}
}