@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);
}
}
}
@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();
}
/** This tests retaining assignments on a cluster restart */
@Test(timeout = 300000)
public void testRetainAssignmentOnRestart() throws Exception {
UTIL.startMiniCluster(2);
while (!UTIL.getMiniHBaseCluster().getMaster().isInitialized()) {
Threads.sleep(1);
}
// Turn off balancer
UTIL.getMiniHBaseCluster().getMaster().getMasterRpcServices().synchronousBalanceSwitch(false);
LOG.info("\n\nCreating tables");
for (byte[] TABLE : TABLES) {
UTIL.createTable(TABLE, FAMILY);
}
for (byte[] TABLE : TABLES) {
UTIL.waitTableEnabled(TABLE);
}
HMaster master = UTIL.getMiniHBaseCluster().getMaster();
UTIL.waitUntilNoRegionsInTransition(120000);
// We don't have to use SnapshotOfRegionAssignmentFromMeta.
// We use it here because AM used to use it to load all user region placements
SnapshotOfRegionAssignmentFromMeta snapshot =
new SnapshotOfRegionAssignmentFromMeta(master.getShortCircuitConnection());
snapshot.initialize();
Map<HRegionInfo, ServerName> regionToRegionServerMap = snapshot.getRegionToRegionServerMap();
MiniHBaseCluster cluster = UTIL.getHBaseCluster();
List<JVMClusterUtil.RegionServerThread> threads = cluster.getLiveRegionServerThreads();
assertEquals(2, threads.size());
int[] rsPorts = new int[3];
for (int i = 0; i < 2; i++) {
rsPorts[i] = threads.get(i).getRegionServer().getServerName().getPort();
}
rsPorts[2] = cluster.getMaster().getServerName().getPort();
for (ServerName serverName : regionToRegionServerMap.values()) {
boolean found = false; // Test only, no need to optimize
for (int k = 0; k < 3 && !found; k++) {
found = serverName.getPort() == rsPorts[k];
}
assertTrue(found);
}
LOG.info("\n\nShutting down HBase cluster");
cluster.shutdown();
cluster.waitUntilShutDown();
LOG.info("\n\nSleeping a bit");
Thread.sleep(2000);
LOG.info("\n\nStarting cluster the second time with the same ports");
try {
cluster.getConf().setInt(ServerManager.WAIT_ON_REGIONSERVERS_MINTOSTART, 4);
master = cluster.startMaster().getMaster();
for (int i = 0; i < 3; i++) {
cluster.getConf().setInt(HConstants.REGIONSERVER_PORT, rsPorts[i]);
cluster.startRegionServer();
}
} finally {
// Reset region server port so as not to conflict with other tests
cluster.getConf().setInt(HConstants.REGIONSERVER_PORT, 0);
cluster.getConf().setInt(ServerManager.WAIT_ON_REGIONSERVERS_MINTOSTART, 2);
}
// Make sure live regionservers are on the same host/port
List<ServerName> localServers = master.getServerManager().getOnlineServersList();
assertEquals(4, localServers.size());
for (int i = 0; i < 3; i++) {
boolean found = false;
for (ServerName serverName : localServers) {
if (serverName.getPort() == rsPorts[i]) {
found = true;
break;
}
}
assertTrue(found);
}
// Wait till master is initialized and all regions are assigned
RegionStates regionStates = master.getAssignmentManager().getRegionStates();
int expectedRegions = regionToRegionServerMap.size() + 1;
while (!master.isInitialized()
|| regionStates.getRegionAssignments().size() != expectedRegions) {
Threads.sleep(100);
}
snapshot = new SnapshotOfRegionAssignmentFromMeta(master.getShortCircuitConnection());
snapshot.initialize();
Map<HRegionInfo, ServerName> newRegionToRegionServerMap = snapshot.getRegionToRegionServerMap();
assertEquals(regionToRegionServerMap.size(), newRegionToRegionServerMap.size());
for (Map.Entry<HRegionInfo, ServerName> entry : newRegionToRegionServerMap.entrySet()) {
if (TableName.NAMESPACE_TABLE_NAME.equals(entry.getKey().getTable())) continue;
ServerName oldServer = regionToRegionServerMap.get(entry.getKey());
ServerName currentServer = entry.getValue();
assertEquals(oldServer.getHostAndPort(), currentServer.getHostAndPort());
assertNotEquals(oldServer.getStartcode(), currentServer.getStartcode());
}
}