/** @throws java.lang.Exception */
@BeforeClass
public static void setUpBeforeClass() throws Exception {
TEST_UTIL.startMiniZKCluster();
Configuration conf = TEST_UTIL.getConfiguration();
conf.setBoolean(HConstants.REPLICATION_ENABLE_KEY, HConstants.REPLICATION_ENABLE_DEFAULT);
admin = new ReplicationAdmin(conf);
}
@Before
public void setup() throws Exception {
TEST_UTIL = new HBaseTestingUtility();
TEST_UTIL.startMiniZKCluster();
conf = TEST_UTIL.getConfiguration();
// Use a different ZK wrapper instance for each tests.
zkw =
new ZooKeeperWatcher(conf, "split-log-manager-tests" + UUID.randomUUID().toString(), null);
ZKUtil.deleteChildrenRecursively(zkw, zkw.baseZNode);
ZKUtil.createAndFailSilent(zkw, zkw.baseZNode);
assertTrue(ZKUtil.checkExists(zkw, zkw.baseZNode) != -1);
LOG.debug(zkw.baseZNode + " created");
ZKUtil.createAndFailSilent(zkw, zkw.splitLogZNode);
assertTrue(ZKUtil.checkExists(zkw, zkw.splitLogZNode) != -1);
LOG.debug(zkw.splitLogZNode + " created");
stopped = false;
resetCounters();
// By default, we let the test manage the error as before, so the server
// does not appear as dead from the master point of view, only from the split log pov.
Mockito.when(sm.isServerOnline(Mockito.any(ServerName.class))).thenReturn(true);
Mockito.when(master.getServerManager()).thenReturn(sm);
to = 4000;
conf.setInt("hbase.splitlog.manager.timeout", to);
conf.setInt("hbase.splitlog.manager.unassigned.timeout", 2 * to);
conf.setInt("hbase.splitlog.manager.timeoutmonitor.period", 100);
to = to + 4 * 100;
}
@Before
public void before() throws Exception {
testingUtility = new HBaseTestingUtility();
testingUtility.startMiniZKCluster();
testingUtility.startMiniCluster(1);
hTable = testingUtility.createTable(TABLE, CF);
}
@Test
public void testRPCException() throws Exception {
HBaseTestingUtility TEST_UTIL = new HBaseTestingUtility();
TEST_UTIL.startMiniZKCluster();
Configuration conf = TEST_UTIL.getConfiguration();
conf.set(HConstants.MASTER_PORT, "0");
HMaster hm = new HMaster(conf);
HServerAddress hma = hm.getMasterAddress();
try {
HMasterInterface inf =
(HMasterInterface)
HBaseRPC.getProxy(
HMasterInterface.class,
HBaseRPCProtocolVersion.versionID,
hma.getInetSocketAddress(),
conf,
100);
inf.isMasterRunning();
fail();
} catch (RemoteException ex) {
assertTrue(
ex.getMessage()
.startsWith(
"org.apache.hadoop.hbase.ipc.ServerNotRunningException: Server is not running yet"));
} catch (Throwable t) {
fail("Unexpected throwable: " + t);
}
}
@BeforeClass
public static void setUpBeforeClass() throws Exception {
// Start up our mini cluster on top of an 0.92 root.dir that has data from
// a 0.92 hbase run -- it has a table with 100 rows in it -- and see if
// we can migrate from 0.92
TEST_UTIL.startMiniZKCluster();
TEST_UTIL.startMiniDFSCluster(1);
Path testdir = TEST_UTIL.getDataTestDir("TestMetaMigrationConvertToPB");
// Untar our test dir.
File untar = untar(new File(testdir.toString()));
// Now copy the untar up into hdfs so when we start hbase, we'll run from it.
Configuration conf = TEST_UTIL.getConfiguration();
FsShell shell = new FsShell(conf);
FileSystem fs = FileSystem.get(conf);
// find where hbase will root itself, so we can copy filesystem there
Path hbaseRootDir = TEST_UTIL.getDefaultRootDirPath();
if (!fs.isDirectory(hbaseRootDir.getParent())) {
// mkdir at first
fs.mkdirs(hbaseRootDir.getParent());
}
doFsCommand(shell, new String[] {"-put", untar.toURI().toString(), hbaseRootDir.toString()});
// windows fix: tgz file has .META. directory renamed as -META- since the original is an illegal
// name under windows. So we rename it back. See
// src/test/data//TestMetaMigrationConvertingToPB.README and
// https://issues.apache.org/jira/browse/HBASE-6821
doFsCommand(
shell,
new String[] {
"-mv",
new Path(hbaseRootDir, "-META-").toString(),
new Path(hbaseRootDir, ".META.").toString()
});
// See whats in minihdfs.
doFsCommand(shell, new String[] {"-lsr", "/"});
TEST_UTIL.startMiniHBaseCluster(1, 1);
// Assert we are running against the copied-up filesystem. The copied-up
// rootdir should have had a table named 'TestTable' in it. Assert it
// present.
HTable t = new HTable(TEST_UTIL.getConfiguration(), TESTTABLE);
ResultScanner scanner = t.getScanner(new Scan());
int count = 0;
while (scanner.next() != null) {
count++;
}
// Assert that we find all 100 rows that are in the data we loaded. If
// so then we must have migrated it from 0.90 to 0.92.
Assert.assertEquals(ROW_COUNT, count);
scanner.close();
t.close();
}
@Before
public void setup() throws Exception {
TEST_UTIL.startMiniZKCluster();
conf = TEST_UTIL.getConfiguration();
zkw = new ZooKeeperWatcher(conf, "split-log-manager-tests", null);
ZKUtil.deleteChildrenRecursively(zkw, zkw.baseZNode);
ZKUtil.createAndFailSilent(zkw, zkw.baseZNode);
assertTrue(ZKUtil.checkExists(zkw, zkw.baseZNode) != -1);
LOG.debug(zkw.baseZNode + " created");
ZKUtil.createAndFailSilent(zkw, zkw.splitLogZNode);
assertTrue(ZKUtil.checkExists(zkw, zkw.splitLogZNode) != -1);
LOG.debug(zkw.splitLogZNode + " created");
stopped = false;
resetCounters();
}
@BeforeClass
public static void setUpBeforeClass() throws Exception {
conf1 = HBaseConfiguration.create();
conf1.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/1");
// smaller block size and capacity to trigger more operations
// and test them
conf1.setInt("hbase.regionserver.hlog.blocksize", 1024 * 20);
conf1.setInt("replication.source.size.capacity", 1024);
conf1.setLong("replication.source.sleepforretries", 100);
conf1.setInt("hbase.regionserver.maxlogs", 10);
conf1.setLong("hbase.master.logcleaner.ttl", 10);
conf1.setBoolean(HConstants.REPLICATION_ENABLE_KEY, true);
conf1.setBoolean("dfs.support.append", true);
conf1.setLong(HConstants.THREAD_WAKE_FREQUENCY, 100);
conf1.setStrings(
CoprocessorHost.USER_REGION_COPROCESSOR_CONF_KEY,
"org.apache.hadoop.hbase.replication.TestMasterReplication$CoprocessorCounter");
utility1 = new HBaseTestingUtility(conf1);
utility1.startMiniZKCluster();
MiniZooKeeperCluster miniZK = utility1.getZkCluster();
new ZooKeeperWatcher(conf1, "cluster1", null, true);
conf2 = new Configuration(conf1);
conf2.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/2");
conf3 = new Configuration(conf1);
conf3.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/3");
utility2 = new HBaseTestingUtility(conf2);
utility2.setZkCluster(miniZK);
new ZooKeeperWatcher(conf2, "cluster3", null, true);
utility3 = new HBaseTestingUtility(conf3);
utility3.setZkCluster(miniZK);
new ZooKeeperWatcher(conf3, "cluster3", null, true);
table = new HTableDescriptor(tableName);
HColumnDescriptor fam = new HColumnDescriptor(famName);
fam.setScope(HConstants.REPLICATION_SCOPE_GLOBAL);
table.addFamily(fam);
fam = new HColumnDescriptor(noRepfamName);
table.addFamily(fam);
}
@SuppressWarnings("resource")
private void startMiniClusters(int numClusters) throws Exception {
Random random = new Random();
utilities = new HBaseTestingUtility[numClusters];
configurations = new Configuration[numClusters];
for (int i = 0; i < numClusters; i++) {
Configuration conf = new Configuration(baseConfiguration);
conf.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/" + i + random.nextInt());
HBaseTestingUtility utility = new HBaseTestingUtility(conf);
if (i == 0) {
utility.startMiniZKCluster();
miniZK = utility.getZkCluster();
} else {
utility.setZkCluster(miniZK);
}
utility.startMiniCluster();
utilities[i] = utility;
configurations[i] = conf;
new ZooKeeperWatcher(conf, "cluster" + i, null, true);
}
}
@Test
public void testRPCException() throws Exception {
HBaseTestingUtility TEST_UTIL = new HBaseTestingUtility();
TEST_UTIL.startMiniZKCluster();
Configuration conf = TEST_UTIL.getConfiguration();
conf.set(HConstants.MASTER_PORT, "0");
HMaster hm = new HMaster(conf);
ServerName sm = hm.getServerName();
InetSocketAddress isa = new InetSocketAddress(sm.getHostname(), sm.getPort());
int i = 0;
// retry the RPC a few times; we have seen SocketTimeoutExceptions if we
// try to connect too soon. Retry on SocketTimeoutException.
while (i < 20) {
try {
MasterMonitorProtocol inf =
(MasterMonitorProtocol)
HBaseClientRPC.getProxy(MasterMonitorProtocol.class, isa, conf, 100 * 10);
inf.isMasterRunning(null, IsMasterRunningRequest.getDefaultInstance());
fail();
} catch (ServiceException ex) {
IOException ie = ProtobufUtil.getRemoteException(ex);
if (!(ie instanceof SocketTimeoutException)) {
if (ie.getMessage()
.startsWith(
"org.apache.hadoop.hbase.ipc.ServerNotRunningYetException: Server is not running yet")) {
return;
}
} else {
System.err.println("Got SocketTimeoutException. Will retry. ");
}
} catch (Throwable t) {
fail("Unexpected throwable: " + t);
}
Thread.sleep(100);
i++;
}
fail();
}
@Before
public void setUp() throws Exception {
File tempDir = Files.createTempDir();
tempDir.deleteOnExit();
htu = HBaseTestingUtility.createLocalHTU();
try {
htu.cleanupTestDir();
htu.startMiniZKCluster();
htu.startMiniHBaseCluster(1, 1);
try {
htu.deleteTable(Bytes.toBytes(tableName));
} catch (Exception e) {
Log.info(" - no table " + tableName + " found");
}
htu.createTable(Bytes.toBytes(tableName), colFam);
dao =
new Hbase1OffsetStore.Builder()
.setHbaseConfiguration(htu.getConfiguration())
.setOffsetTable(tableName)
.build();
} catch (Exception e1) {
throw new RuntimeException(e1);
}
KOM =
new KafkaOffsetManager.Builder()
.setOffsetManager(dao)
.setKafkaBrokerList("localhost:" + kafkaRule.kafkaBrokerPort())
.setGroupID(testGroupID)
.setTopic(testTopicName)
.build();
}
/** @throws java.lang.Exception */
@BeforeClass
public static void setUpBeforeClass() throws Exception {
conf1.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/1");
// smaller log roll size to trigger more events
conf1.setFloat("hbase.regionserver.logroll.multiplier", 0.0003f);
conf1.setInt("replication.source.size.capacity", 10240);
conf1.setLong("replication.source.sleepforretries", 100);
conf1.setInt("hbase.regionserver.maxlogs", 10);
conf1.setLong("hbase.master.logcleaner.ttl", 10);
conf1.setInt("zookeeper.recovery.retry", 1);
conf1.setInt("zookeeper.recovery.retry.intervalmill", 10);
conf1.setBoolean(HConstants.REPLICATION_ENABLE_KEY, HConstants.REPLICATION_ENABLE_DEFAULT);
conf1.setBoolean("dfs.support.append", true);
conf1.setLong(HConstants.THREAD_WAKE_FREQUENCY, 100);
conf1.setInt("replication.stats.thread.period.seconds", 5);
conf1.setBoolean("hbase.tests.use.shortcircuit.reads", false);
utility1 = new HBaseTestingUtility(conf1);
utility1.startMiniZKCluster();
MiniZooKeeperCluster miniZK = utility1.getZkCluster();
// Have to reget conf1 in case zk cluster location different
// than default
conf1 = utility1.getConfiguration();
zkw1 = new ZooKeeperWatcher(conf1, "cluster1", null, true);
admin = new ReplicationAdmin(conf1);
LOG.info("Setup first Zk");
// Base conf2 on conf1 so it gets the right zk cluster.
conf2 = HBaseConfiguration.create(conf1);
conf2.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/2");
conf2.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, 6);
conf2.setBoolean(HConstants.REPLICATION_ENABLE_KEY, HConstants.REPLICATION_ENABLE_DEFAULT);
conf2.setBoolean("dfs.support.append", true);
conf2.setBoolean("hbase.tests.use.shortcircuit.reads", false);
utility2 = new HBaseTestingUtility(conf2);
utility2.setZkCluster(miniZK);
zkw2 = new ZooKeeperWatcher(conf2, "cluster2", null, true);
admin.addPeer("2", utility2.getClusterKey());
LOG.info("Setup second Zk");
CONF_WITH_LOCALFS = HBaseConfiguration.create(conf1);
utility1.startMiniCluster(2);
utility2.startMiniCluster(2);
HTableDescriptor table = new HTableDescriptor(TableName.valueOf(tableName));
HColumnDescriptor fam = new HColumnDescriptor(famName);
fam.setMaxVersions(3);
fam.setScope(HConstants.REPLICATION_SCOPE_GLOBAL);
table.addFamily(fam);
fam = new HColumnDescriptor(noRepfamName);
table.addFamily(fam);
HBaseAdmin admin1 = new HBaseAdmin(conf1);
HBaseAdmin admin2 = new HBaseAdmin(conf2);
admin1.createTable(table, HBaseTestingUtility.KEYS_FOR_HBA_CREATE_TABLE);
admin2.createTable(table, HBaseTestingUtility.KEYS_FOR_HBA_CREATE_TABLE);
htable1 = new HTable(conf1, tableName);
htable1.setWriteBufferSize(1024);
htable2 = new HTable(conf2, tableName);
}
@BeforeClass
public static void beforeClass() throws Exception {
// Set this down so tests run quicker
UTIL.getConfiguration().setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, 3);
UTIL.startMiniZKCluster();
}
@Test
public void testReducerNumEstimation() throws Exception {
// Skip the test for Tez. Tez use a different mechanism.
// Equivalent test is in TestTezAutoParallelism
Assume.assumeTrue("Skip this test for TEZ", Util.isMapredExecType(cluster.getExecType()));
// use the estimation
Configuration conf = HBaseConfiguration.create(new Configuration());
HBaseTestingUtility util = new HBaseTestingUtility(conf);
int clientPort = util.startMiniZKCluster().getClientPort();
util.startMiniHBaseCluster(1, 1);
String query = "a = load '/passwd';" + "b = group a by $0;" + "store b into 'output';";
PigServer ps = new PigServer(cluster.getExecType(), cluster.getProperties());
PhysicalPlan pp = Util.buildPp(ps, query);
MROperPlan mrPlan = Util.buildMRPlan(pp, pc);
pc.getConf().setProperty("pig.exec.reducers.bytes.per.reducer", "100");
pc.getConf().setProperty("pig.exec.reducers.max", "10");
pc.getConf().setProperty(HConstants.ZOOKEEPER_CLIENT_PORT, Integer.toString(clientPort));
ConfigurationValidator.validatePigProperties(pc.getProperties());
conf = ConfigurationUtil.toConfiguration(pc.getProperties());
JobControlCompiler jcc = new JobControlCompiler(pc, conf);
JobControl jc = jcc.compile(mrPlan, "Test");
Job job = jc.getWaitingJobs().get(0);
long reducer =
Math.min(
(long) Math.ceil(new File("test/org/apache/pig/test/data/passwd").length() / 100.0),
10);
Util.assertParallelValues(-1, -1, reducer, reducer, job.getJobConf());
// use the PARALLEL key word, it will override the estimated reducer number
query = "a = load '/passwd';" + "b = group a by $0 PARALLEL 2;" + "store b into 'output';";
pp = Util.buildPp(ps, query);
mrPlan = Util.buildMRPlan(pp, pc);
pc.getConf().setProperty("pig.exec.reducers.bytes.per.reducer", "100");
pc.getConf().setProperty("pig.exec.reducers.max", "10");
ConfigurationValidator.validatePigProperties(pc.getProperties());
conf = ConfigurationUtil.toConfiguration(pc.getProperties());
jcc = new JobControlCompiler(pc, conf);
jc = jcc.compile(mrPlan, "Test");
job = jc.getWaitingJobs().get(0);
Util.assertParallelValues(-1, 2, -1, 2, job.getJobConf());
final byte[] COLUMNFAMILY = Bytes.toBytes("pig");
util.createTable(Bytes.toBytesBinary("test_table"), COLUMNFAMILY);
// the estimation won't take effect when it apply to non-dfs or the files doesn't exist, such as
// hbase
query =
"a = load 'hbase://test_table' using org.apache.pig.backend.hadoop.hbase.HBaseStorage('c:f1 c:f2');"
+ "b = group a by $0 ;"
+ "store b into 'output';";
pp = Util.buildPp(ps, query);
mrPlan = Util.buildMRPlan(pp, pc);
pc.getConf().setProperty("pig.exec.reducers.bytes.per.reducer", "100");
pc.getConf().setProperty("pig.exec.reducers.max", "10");
ConfigurationValidator.validatePigProperties(pc.getProperties());
conf = ConfigurationUtil.toConfiguration(pc.getProperties());
jcc = new JobControlCompiler(pc, conf);
jc = jcc.compile(mrPlan, "Test");
job = jc.getWaitingJobs().get(0);
Util.assertParallelValues(-1, -1, 1, 1, job.getJobConf());
util.deleteTable(Bytes.toBytesBinary("test_table"));
// In HBase 0.90.1 and above we can use util.shutdownMiniHBaseCluster()
// here instead.
MiniHBaseCluster hbc = util.getHBaseCluster();
if (hbc != null) {
hbc.shutdown();
hbc.join();
}
util.shutdownMiniZKCluster();
}
@Before
public void before() throws Exception {
UTIL.startMiniZKCluster();
}
@BeforeClass
public static void setUpBeforeClass() throws Exception {
conf1.setInt("hfile.format.version", 3);
conf1.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/1");
conf1.setInt("replication.source.size.capacity", 10240);
conf1.setLong("replication.source.sleepforretries", 100);
conf1.setInt("hbase.regionserver.maxlogs", 10);
conf1.setLong("hbase.master.logcleaner.ttl", 10);
conf1.setInt("zookeeper.recovery.retry", 1);
conf1.setInt("zookeeper.recovery.retry.intervalmill", 10);
conf1.setBoolean("dfs.support.append", true);
conf1.setLong(HConstants.THREAD_WAKE_FREQUENCY, 100);
conf1.setInt("replication.stats.thread.period.seconds", 5);
conf1.setBoolean("hbase.tests.use.shortcircuit.reads", false);
conf1.setStrings(HConstants.REPLICATION_CODEC_CONF_KEY, KeyValueCodecWithTags.class.getName());
conf1.setStrings(
CoprocessorHost.USER_REGION_COPROCESSOR_CONF_KEY,
TestCoprocessorForTagsAtSource.class.getName());
utility1 = new HBaseTestingUtility(conf1);
utility1.startMiniZKCluster();
MiniZooKeeperCluster miniZK = utility1.getZkCluster();
// Have to reget conf1 in case zk cluster location different
// than default
conf1 = utility1.getConfiguration();
replicationAdmin = new ReplicationAdmin(conf1);
LOG.info("Setup first Zk");
// Base conf2 on conf1 so it gets the right zk cluster.
conf2 = HBaseConfiguration.create(conf1);
conf2.setInt("hfile.format.version", 3);
conf2.set(HConstants.ZOOKEEPER_ZNODE_PARENT, "/2");
conf2.setInt(HConstants.HBASE_CLIENT_RETRIES_NUMBER, 6);
conf2.setBoolean("dfs.support.append", true);
conf2.setBoolean("hbase.tests.use.shortcircuit.reads", false);
conf2.setStrings(HConstants.REPLICATION_CODEC_CONF_KEY, KeyValueCodecWithTags.class.getName());
conf2.setStrings(
CoprocessorHost.USER_REGION_COPROCESSOR_CONF_KEY,
TestCoprocessorForTagsAtSink.class.getName());
utility2 = new HBaseTestingUtility(conf2);
utility2.setZkCluster(miniZK);
replicationAdmin.addPeer("2", utility2.getClusterKey());
LOG.info("Setup second Zk");
utility1.startMiniCluster(2);
utility2.startMiniCluster(2);
HTableDescriptor table = new HTableDescriptor(TABLE_NAME);
HColumnDescriptor fam = new HColumnDescriptor(FAMILY);
fam.setMaxVersions(3);
fam.setScope(HConstants.REPLICATION_SCOPE_GLOBAL);
table.addFamily(fam);
try (Connection conn = ConnectionFactory.createConnection(conf1);
Admin admin = conn.getAdmin()) {
admin.createTable(table, HBaseTestingUtility.KEYS_FOR_HBA_CREATE_TABLE);
}
try (Connection conn = ConnectionFactory.createConnection(conf2);
Admin admin = conn.getAdmin()) {
admin.createTable(table, HBaseTestingUtility.KEYS_FOR_HBA_CREATE_TABLE);
}
htable1 = utility1.getConnection().getTable(TABLE_NAME);
htable2 = utility2.getConnection().getTable(TABLE_NAME);
}
/** @throws java.lang.Exception */
@BeforeClass
public static void setUpBeforeClass() throws Exception {
TEST_UTIL.startMiniZKCluster();
}
@BeforeClass
public static void before() throws Exception {
HTU.startMiniZKCluster();
}
@Test
public void testHostRank() throws Exception {
if (System.getProperty("prop.mapred.job.tracker") != null) {
if (LOG.isInfoEnabled())
LOG.info("testHBaseInputOutput: Ignore this test if not local mode.");
return;
}
File jarTest = new File(System.getProperty("prop.jarLocation"));
if (!jarTest.exists()) {
fail(
"Could not find Giraph jar at "
+ "location specified by 'prop.jarLocation'. "
+ "Make sure you built the main Giraph artifact?.");
}
MiniHBaseCluster cluster = null;
MiniZooKeeperCluster zkCluster = null;
FileSystem fs = null;
try {
// using the restart method allows us to avoid having the hbase
// root directory overwritten by /home/$username
zkCluster = testUtil.startMiniZKCluster();
testUtil.restartHBaseCluster(2);
cluster = testUtil.getMiniHBaseCluster();
final byte[] OL_BYTES = Bytes.toBytes("ol");
final byte[] S_BYTES = Bytes.toBytes("s");
final byte[] METADATA_BYTES = Bytes.toBytes("mtdt");
final byte[] HR_BYTES = Bytes.toBytes("_hr_");
final byte[] TAB = Bytes.toBytes(TABLE_NAME);
Configuration conf = cluster.getConfiguration();
HTableDescriptor desc = new HTableDescriptor(TAB);
desc.addFamily(new HColumnDescriptor(OL_BYTES));
desc.addFamily(new HColumnDescriptor(S_BYTES));
desc.addFamily(new HColumnDescriptor(METADATA_BYTES));
HBaseAdmin hbaseAdmin = new HBaseAdmin(conf);
if (hbaseAdmin.isTableAvailable(TABLE_NAME)) {
hbaseAdmin.disableTable(TABLE_NAME);
hbaseAdmin.deleteTable(TABLE_NAME);
}
hbaseAdmin.createTable(desc);
/**
* Enter the initial data (a,b), (b,c), (a,c) a = 1.0 - google b = 1.0 - yahoo c = 1.0 - bing
*/
HTable table = new HTable(conf, TABLE_NAME);
Put p1 = new Put(Bytes.toBytes("com.google.www"));
p1.add(OL_BYTES, Bytes.toBytes("www.yahoo.com"), Bytes.toBytes("ab"));
Put p2 = new Put(Bytes.toBytes("com.google.www"));
p2.add(OL_BYTES, Bytes.toBytes("www.bing.com"), Bytes.toBytes("ac"));
p2.add(OL_BYTES, Bytes.toBytes("www.bing.com"), Bytes.toBytes("invalid1"));
p2.add(OL_BYTES, Bytes.toBytes("www.google.com"), Bytes.toBytes("invalid2"));
Put p3 = new Put(Bytes.toBytes("com.yahoo.www"));
p3.add(OL_BYTES, Bytes.toBytes("www.bing.com"), Bytes.toBytes("bc"));
// p3.add(OL_BYTES, Bytes.toBytes(""), Bytes.toBytes("invalid4"));
Put p4 = new Put(Bytes.toBytes("com.bing.www"));
// TODO: Handle below case. use apache isValid method.
p4.add(OL_BYTES, Bytes.toBytes("http://invalidurl"), Bytes.toBytes("invalid5"));
p4.add(S_BYTES, S_BYTES, Bytes.toBytes(10.0d));
Put p5 = new Put(Bytes.toBytes("dummy"));
p5.add(S_BYTES, S_BYTES, Bytes.toBytes(10.0d));
table.put(p1);
table.put(p2);
table.put(p3);
table.put(p4);
table.put(p5);
// Set Giraph configuration
// now operate over HBase using Vertex I/O formats
conf.set(TableInputFormat.INPUT_TABLE, TABLE_NAME);
conf.set(TableOutputFormat.OUTPUT_TABLE, TABLE_NAME);
// Start the giraph job
GiraphJob giraphJob = new GiraphJob(conf, BspCase.getCallingMethodName());
GiraphConfiguration giraphConf = giraphJob.getConfiguration();
giraphConf.setZooKeeperConfiguration(cluster.getMaster().getZooKeeper().getQuorum());
setupConfiguration(giraphJob);
giraphConf.setComputationClass(LinkRankComputation.class);
giraphConf.setMasterComputeClass(LinkRankVertexMasterCompute.class);
giraphConf.setOutEdgesClass(ByteArrayEdges.class);
giraphConf.setVertexInputFormatClass(Nutch2HostInputFormat.class);
giraphConf.setVertexOutputFormatClass(Nutch2HostOutputFormat.class);
giraphConf.setInt("giraph.linkRank.superstepCount", 10);
giraphConf.setInt("giraph.linkRank.scale", 10);
giraphConf.set("giraph.linkRank.family", "mtdt");
giraphConf.set("giraph.linkRank.qualifier", "_hr_");
giraphConf.setVertexInputFilterClass(HostRankVertexFilter.class);
assertTrue(giraphJob.run(true));
if (LOG.isInfoEnabled()) LOG.info("Giraph job successful. Checking output qualifier.");
/** Check the results * */
Result result;
String key;
byte[] calculatedScoreByte;
HashMap expectedValues = new HashMap<String, Double>();
expectedValues.put("com.google.www", 1.3515060339386287d);
expectedValues.put("com.yahoo.www", 4.144902009567587d);
expectedValues.put("com.bing.www", 9.063893290511482d);
for (Object keyObject : expectedValues.keySet()) {
key = keyObject.toString();
result = table.get(new Get(key.getBytes()));
calculatedScoreByte = result.getValue(METADATA_BYTES, HR_BYTES);
assertNotNull(calculatedScoreByte);
assertTrue(calculatedScoreByte.length > 0);
Assert.assertEquals(
"Scores are not the same",
(Double) expectedValues.get(key),
Bytes.toDouble(calculatedScoreByte),
DELTA);
}
} finally {
if (cluster != null) {
cluster.shutdown();
}
if (zkCluster != null) {
zkCluster.shutdown();
}
// clean test files
if (fs != null) {
fs.delete(hbaseRootdir);
}
}
}