Examples with CompactionRequest org.apache.hadoop.hive.metastore.api.CompactionRequest used on opensource projects

Example 21 with CompactionRequest

use of org.apache.hadoop.hive.metastore.api.CompactionRequest in project hive by apache.

the class TestCompactor method majorCompactWhileStreamingForSplitUpdate.

@Test
public void majorCompactWhileStreamingForSplitUpdate() throws Exception {
    String dbName = "default";
    String tblName = "cws";
    List<String> colNames = Arrays.asList("a", "b");
    String columnNamesProperty = "a,b";
    String columnTypesProperty = "int:string";
    executeStatementOnDriver("drop table if exists " + tblName, driver);
    executeStatementOnDriver("CREATE TABLE " + tblName + "(a INT, b STRING) " + //currently ACID requires table to be bucketed
    " CLUSTERED BY(a) INTO 2 BUCKETS" + " STORED AS ORC  TBLPROPERTIES ('transactional'='true', " + "'transactional_properties'='default') ", // this turns on split-update U=D+I
    driver);
    HiveEndPoint endPt = new HiveEndPoint(null, dbName, tblName, null);
    DelimitedInputWriter writer = new DelimitedInputWriter(new String[] { "a", "b" }, ",", endPt);
    StreamingConnection connection = endPt.newConnection(false, "UT_" + Thread.currentThread().getName());
    try {
        // Write a couple of batches
        for (int i = 0; i < 2; i++) {
            writeBatch(connection, writer, false);
        }
        // Start a third batch, but don't close it.
        writeBatch(connection, writer, true);
        // Now, compact
        TxnStore txnHandler = TxnUtils.getTxnStore(conf);
        txnHandler.compact(new CompactionRequest(dbName, tblName, CompactionType.MAJOR));
        Worker t = new Worker();
        t.setThreadId((int) t.getId());
        t.setHiveConf(conf);
        AtomicBoolean stop = new AtomicBoolean(true);
        AtomicBoolean looped = new AtomicBoolean();
        t.init(stop, looped);
        t.run();
        // Find the location of the table
        IMetaStoreClient msClient = new HiveMetaStoreClient(conf);
        Table table = msClient.getTable(dbName, tblName);
        FileSystem fs = FileSystem.get(conf);
        FileStatus[] stat = fs.listStatus(new Path(table.getSd().getLocation()), AcidUtils.baseFileFilter);
        if (1 != stat.length) {
            Assert.fail("Expecting 1 file \"base_0000004\" and found " + stat.length + " files " + Arrays.toString(stat));
        }
        String name = stat[0].getPath().getName();
        Assert.assertEquals(name, "base_0000004");
        checkExpectedTxnsPresent(stat[0].getPath(), null, columnNamesProperty, columnTypesProperty, 0, 1L, 4L);
    } finally {
        connection.close();
    }
}

Example 22 with CompactionRequest

use of org.apache.hadoop.hive.metastore.api.CompactionRequest in project hive by apache.

the class TestCompactor method testMinorCompactionForSplitUpdateWithOnlyInserts.

@Test
public void testMinorCompactionForSplitUpdateWithOnlyInserts() throws Exception {
    String agentInfo = "UT_" + Thread.currentThread().getName();
    String dbName = "default";
    String tblName = "cws";
    List<String> colNames = Arrays.asList("a", "b");
    String columnNamesProperty = "a,b";
    String columnTypesProperty = "int:string";
    executeStatementOnDriver("drop table if exists " + tblName, driver);
    executeStatementOnDriver("CREATE TABLE " + tblName + "(a INT, b STRING) " + //currently ACID requires table to be bucketed
    " CLUSTERED BY(a) INTO 1 BUCKETS" + " STORED AS ORC  TBLPROPERTIES ('transactional'='true'," + "'transactional_properties'='default')", driver);
    // Insert some data -> this will generate only insert deltas and no delete deltas: delta_1_1
    executeStatementOnDriver("INSERT INTO " + tblName + "(a,b) VALUES(1, 'foo')", driver);
    // Insert some data -> this will again generate only insert deltas and no delete deltas: delta_2_2
    executeStatementOnDriver("INSERT INTO " + tblName + "(a,b) VALUES(2, 'bar')", driver);
    // Now, compact
    // One important thing to note in this test is that minor compaction always produces
    // delta_x_y and a counterpart delete_delta_x_y, even when there are no delete_delta events.
    // Such a choice has been made to simplify processing of AcidUtils.getAcidState().
    TxnStore txnHandler = TxnUtils.getTxnStore(conf);
    txnHandler.compact(new CompactionRequest(dbName, tblName, CompactionType.MINOR));
    Worker t = new Worker();
    t.setThreadId((int) t.getId());
    t.setHiveConf(conf);
    AtomicBoolean stop = new AtomicBoolean(true);
    AtomicBoolean looped = new AtomicBoolean();
    t.init(stop, looped);
    t.run();
    // Find the location of the table
    IMetaStoreClient msClient = new HiveMetaStoreClient(conf);
    Table table = msClient.getTable(dbName, tblName);
    FileSystem fs = FileSystem.get(conf);
    // Verify that we have got correct set of deltas.
    FileStatus[] stat = fs.listStatus(new Path(table.getSd().getLocation()), AcidUtils.deltaFileFilter);
    String[] deltas = new String[stat.length];
    Path minorCompactedDelta = null;
    for (int i = 0; i < deltas.length; i++) {
        deltas[i] = stat[i].getPath().getName();
        if (deltas[i].equals("delta_0000001_0000002")) {
            minorCompactedDelta = stat[i].getPath();
        }
    }
    Arrays.sort(deltas);
    String[] expectedDeltas = new String[] { "delta_0000001_0000001_0000", "delta_0000001_0000002", "delta_0000002_0000002_0000" };
    if (!Arrays.deepEquals(expectedDeltas, deltas)) {
        Assert.fail("Expected: " + Arrays.toString(expectedDeltas) + ", found: " + Arrays.toString(deltas));
    }
    checkExpectedTxnsPresent(null, new Path[] { minorCompactedDelta }, columnNamesProperty, columnTypesProperty, 0, 1L, 2L);
    // Verify that we have got correct set of delete_deltas.
    FileStatus[] deleteDeltaStat = fs.listStatus(new Path(table.getSd().getLocation()), AcidUtils.deleteEventDeltaDirFilter);
    String[] deleteDeltas = new String[deleteDeltaStat.length];
    Path minorCompactedDeleteDelta = null;
    for (int i = 0; i < deleteDeltas.length; i++) {
        deleteDeltas[i] = deleteDeltaStat[i].getPath().getName();
        if (deleteDeltas[i].equals("delete_delta_0000001_0000002")) {
            minorCompactedDeleteDelta = deleteDeltaStat[i].getPath();
        }
    }
    Arrays.sort(deleteDeltas);
    String[] expectedDeleteDeltas = new String[] { "delete_delta_0000001_0000002" };
    if (!Arrays.deepEquals(expectedDeleteDeltas, deleteDeltas)) {
        Assert.fail("Expected: " + Arrays.toString(expectedDeleteDeltas) + ", found: " + Arrays.toString(deleteDeltas));
    }
    // There should be no rows in the delete_delta because there have been no delete events.
    checkExpectedTxnsPresent(null, new Path[] { minorCompactedDeleteDelta }, columnNamesProperty, columnTypesProperty, 0, 0L, 0L);
}

Example 23 with CompactionRequest

use of org.apache.hadoop.hive.metastore.api.CompactionRequest in project hive by apache.

the class TestCompactionTxnHandler method testFindNextToCompact.

@Test
public void testFindNextToCompact() throws Exception {
    CompactionRequest rqst = new CompactionRequest("foo", "bar", CompactionType.MINOR);
    rqst.setPartitionname("ds=today");
    txnHandler.compact(rqst);
    long now = System.currentTimeMillis();
    CompactionInfo ci = txnHandler.findNextToCompact("fred");
    assertNotNull(ci);
    assertEquals("foo", ci.dbname);
    assertEquals("bar", ci.tableName);
    assertEquals("ds=today", ci.partName);
    assertEquals(CompactionType.MINOR, ci.type);
    assertNull(ci.runAs);
    assertNull(txnHandler.findNextToCompact("fred"));
    txnHandler.setRunAs(ci.id, "bob");
    ShowCompactResponse rsp = txnHandler.showCompact(new ShowCompactRequest());
    List<ShowCompactResponseElement> compacts = rsp.getCompacts();
    assertEquals(1, compacts.size());
    ShowCompactResponseElement c = compacts.get(0);
    assertEquals("foo", c.getDbname());
    assertEquals("bar", c.getTablename());
    assertEquals("ds=today", c.getPartitionname());
    assertEquals(CompactionType.MINOR, c.getType());
    assertEquals("working", c.getState());
    assertTrue(c.getStart() - 5000 < now && c.getStart() + 5000 > now);
    assertEquals("fred", c.getWorkerid());
    assertEquals("bob", c.getRunAs());
}

Example 24 with CompactionRequest

use of org.apache.hadoop.hive.metastore.api.CompactionRequest in project hive by apache.

the class TestCompactionTxnHandler method testRevokeFromLocalWorkers.

@Test
public void testRevokeFromLocalWorkers() throws Exception {
    CompactionRequest rqst = new CompactionRequest("foo", "bar", CompactionType.MINOR);
    txnHandler.compact(rqst);
    rqst = new CompactionRequest("foo", "baz", CompactionType.MINOR);
    txnHandler.compact(rqst);
    rqst = new CompactionRequest("foo", "bazzoo", CompactionType.MINOR);
    txnHandler.compact(rqst);
    assertNotNull(txnHandler.findNextToCompact("fred-193892"));
    assertNotNull(txnHandler.findNextToCompact("bob-193892"));
    assertNotNull(txnHandler.findNextToCompact("fred-193893"));
    txnHandler.revokeFromLocalWorkers("fred");
    ShowCompactResponse rsp = txnHandler.showCompact(new ShowCompactRequest());
    List<ShowCompactResponseElement> compacts = rsp.getCompacts();
    assertEquals(3, compacts.size());
    boolean sawWorkingBob = false;
    int initiatedCount = 0;
    for (ShowCompactResponseElement c : compacts) {
        if (c.getState().equals("working")) {
            assertEquals("bob-193892", c.getWorkerid());
            sawWorkingBob = true;
        } else if (c.getState().equals("initiated")) {
            initiatedCount++;
        } else {
            fail("Unexpected state");
        }
    }
    assertTrue(sawWorkingBob);
    assertEquals(2, initiatedCount);
}

Example 25 with CompactionRequest

use of org.apache.hadoop.hive.metastore.api.CompactionRequest in project hive by apache.

the class TestCompactionTxnHandler method testRevokeTimedOutWorkers.

@Test
public void testRevokeTimedOutWorkers() throws Exception {
    CompactionRequest rqst = new CompactionRequest("foo", "bar", CompactionType.MINOR);
    txnHandler.compact(rqst);
    rqst = new CompactionRequest("foo", "baz", CompactionType.MINOR);
    txnHandler.compact(rqst);
    assertNotNull(txnHandler.findNextToCompact("fred-193892"));
    Thread.sleep(200);
    assertNotNull(txnHandler.findNextToCompact("fred-193892"));
    txnHandler.revokeTimedoutWorkers(100);
    ShowCompactResponse rsp = txnHandler.showCompact(new ShowCompactRequest());
    List<ShowCompactResponseElement> compacts = rsp.getCompacts();
    assertEquals(2, compacts.size());
    boolean sawWorking = false, sawInitiated = false;
    for (ShowCompactResponseElement c : compacts) {
        if (c.getState().equals("working"))
            sawWorking = true;
        else if (c.getState().equals("initiated"))
            sawInitiated = true;
        else
            fail("Unexpected state");
    }
    assertTrue(sawWorking);
    assertTrue(sawInitiated);
}