Examples with RootTabletStateStore org.apache.accumulo.server.master.state.RootTabletStateStore used on opensource projects

Example 1 with RootTabletStateStore

use of org.apache.accumulo.server.master.state.RootTabletStateStore in project accumulo by apache.

the class Master method run.

public void run() throws IOException, InterruptedException, KeeperException {
    final String zroot = ZooUtil.getRoot(getInstance());
    // ACCUMULO-4424 Put up the Thrift servers before getting the lock as a sign of process health when a hot-standby
    // 
    // Start the Master's Client service
    clientHandler = new MasterClientServiceHandler(this);
    // Ensure that calls before the master gets the lock fail
    Iface haProxy = HighlyAvailableServiceWrapper.service(clientHandler, this);
    Iface rpcProxy = RpcWrapper.service(haProxy);
    final Processor<Iface> processor;
    if (ThriftServerType.SASL == getThriftServerType()) {
        Iface tcredsProxy = TCredentialsUpdatingWrapper.service(rpcProxy, clientHandler.getClass(), getConfiguration());
        processor = new Processor<>(tcredsProxy);
    } else {
        processor = new Processor<>(rpcProxy);
    }
    ServerAddress sa = TServerUtils.startServer(this, hostname, Property.MASTER_CLIENTPORT, processor, "Master", "Master Client Service Handler", null, Property.MASTER_MINTHREADS, Property.MASTER_THREADCHECK, Property.GENERAL_MAX_MESSAGE_SIZE);
    clientService = sa.server;
    log.info("Started Master client service at {}", sa.address);
    // Start the replication coordinator which assigns tservers to service replication requests
    MasterReplicationCoordinator impl = new MasterReplicationCoordinator(this);
    ReplicationCoordinator.Iface haReplicationProxy = HighlyAvailableServiceWrapper.service(impl, this);
    ReplicationCoordinator.Processor<ReplicationCoordinator.Iface> replicationCoordinatorProcessor = new ReplicationCoordinator.Processor<>(RpcWrapper.service(haReplicationProxy));
    ServerAddress replAddress = TServerUtils.startServer(this, hostname, Property.MASTER_REPLICATION_COORDINATOR_PORT, replicationCoordinatorProcessor, "Master Replication Coordinator", "Replication Coordinator", null, Property.MASTER_REPLICATION_COORDINATOR_MINTHREADS, Property.MASTER_REPLICATION_COORDINATOR_THREADCHECK, Property.GENERAL_MAX_MESSAGE_SIZE);
    log.info("Started replication coordinator service at " + replAddress.address);
    // block until we can obtain the ZK lock for the master
    getMasterLock(zroot + Constants.ZMASTER_LOCK);
    recoveryManager = new RecoveryManager(this);
    TableManager.getInstance().addObserver(this);
    StatusThread statusThread = new StatusThread();
    statusThread.start();
    MigrationCleanupThread migrationCleanupThread = new MigrationCleanupThread();
    migrationCleanupThread.start();
    tserverSet.startListeningForTabletServerChanges();
    ZooReaderWriter zReaderWriter = ZooReaderWriter.getInstance();
    zReaderWriter.getChildren(zroot + Constants.ZRECOVERY, new Watcher() {

        @Override
        public void process(WatchedEvent event) {
            nextEvent.event("Noticed recovery changes", event.getType());
            try {
                // watcher only fires once, add it back
                ZooReaderWriter.getInstance().getChildren(zroot + Constants.ZRECOVERY, this);
            } catch (Exception e) {
                log.error("Failed to add log recovery watcher back", e);
            }
        }
    });
    watchers.add(new TabletGroupWatcher(this, new MetaDataStateStore(this, this), null) {

        @Override
        boolean canSuspendTablets() {
            // Always allow user data tablets to enter suspended state.
            return true;
        }
    });
    watchers.add(new TabletGroupWatcher(this, new RootTabletStateStore(this, this), watchers.get(0)) {

        @Override
        boolean canSuspendTablets() {
            // be immediately reassigned, even if there's a global table.suspension.duration setting.
            return getConfiguration().getBoolean(Property.MASTER_METADATA_SUSPENDABLE);
        }
    });
    watchers.add(new TabletGroupWatcher(this, new ZooTabletStateStore(new ZooStore(zroot)), watchers.get(1)) {

        @Override
        boolean canSuspendTablets() {
            // Never allow root tablet to enter suspended state.
            return false;
        }
    });
    for (TabletGroupWatcher watcher : watchers) {
        watcher.start();
    }
    // Once we are sure the upgrade is complete, we can safely allow fate use.
    waitForMetadataUpgrade.await();
    try {
        final AgeOffStore<Master> store = new AgeOffStore<>(new org.apache.accumulo.fate.ZooStore<Master>(ZooUtil.getRoot(getInstance()) + Constants.ZFATE, ZooReaderWriter.getInstance()), 1000 * 60 * 60 * 8);
        int threads = getConfiguration().getCount(Property.MASTER_FATE_THREADPOOL_SIZE);
        fate = new Fate<>(this, store);
        fate.startTransactionRunners(threads);
        SimpleTimer.getInstance(getConfiguration()).schedule(new Runnable() {

            @Override
            public void run() {
                store.ageOff();
            }
        }, 63000, 63000);
    } catch (KeeperException | InterruptedException e) {
        throw new IOException(e);
    }
    ZooKeeperInitialization.ensureZooKeeperInitialized(zReaderWriter, zroot);
    // the master client service.
    if (null != authenticationTokenKeyManager && null != keyDistributor) {
        log.info("Starting delegation-token key manager");
        keyDistributor.initialize();
        authenticationTokenKeyManager.start();
        boolean logged = false;
        while (!authenticationTokenKeyManager.isInitialized()) {
            // Print out a status message when we start waiting for the key manager to get initialized
            if (!logged) {
                log.info("Waiting for AuthenticationTokenKeyManager to be initialized");
                logged = true;
            }
            sleepUninterruptibly(200, TimeUnit.MILLISECONDS);
        }
        // And log when we are initialized
        log.info("AuthenticationTokenSecretManager is initialized");
    }
    String address = sa.address.toString();
    log.info("Setting master lock data to {}", address);
    masterLock.replaceLockData(address.getBytes());
    while (!clientService.isServing()) {
        sleepUninterruptibly(100, TimeUnit.MILLISECONDS);
    }
    // Start the daemon to scan the replication table and make units of work
    replicationWorkDriver = new ReplicationDriver(this);
    replicationWorkDriver.start();
    // Start the daemon to assign work to tservers to replicate to our peers
    try {
        replicationWorkAssigner = new WorkDriver(this);
    } catch (AccumuloException | AccumuloSecurityException e) {
        log.error("Caught exception trying to initialize replication WorkDriver", e);
        throw new RuntimeException(e);
    }
    replicationWorkAssigner.start();
    // Advertise that port we used so peers don't have to be told what it is
    ZooReaderWriter.getInstance().putPersistentData(ZooUtil.getRoot(getInstance()) + Constants.ZMASTER_REPLICATION_COORDINATOR_ADDR, replAddress.address.toString().getBytes(UTF_8), NodeExistsPolicy.OVERWRITE);
    // Register replication metrics
    MasterMetricsFactory factory = new MasterMetricsFactory(getConfiguration(), this);
    Metrics replicationMetrics = factory.createReplicationMetrics();
    try {
        replicationMetrics.register();
    } catch (Exception e) {
        log.error("Failed to register replication metrics", e);
    }
    // The master is fully initialized. Clients are allowed to connect now.
    masterInitialized.set(true);
    while (clientService.isServing()) {
        sleepUninterruptibly(500, TimeUnit.MILLISECONDS);
    }
    log.info("Shutting down fate.");
    fate.shutdown();
    log.info("Shutting down timekeeping.");
    timeKeeper.shutdown();
    final long deadline = System.currentTimeMillis() + MAX_CLEANUP_WAIT_TIME;
    statusThread.join(remaining(deadline));
    replicationWorkAssigner.join(remaining(deadline));
    replicationWorkDriver.join(remaining(deadline));
    replAddress.server.stop();
    // Signal that we want it to stop, and wait for it to do so.
    if (authenticationTokenKeyManager != null) {
        authenticationTokenKeyManager.gracefulStop();
        authenticationTokenKeyManager.join(remaining(deadline));
    }
    // don't stop
    for (TabletGroupWatcher watcher : watchers) {
        watcher.join(remaining(deadline));
    }
    log.info("exiting");
}

Example 2 with RootTabletStateStore

use of org.apache.accumulo.server.master.state.RootTabletStateStore in project accumulo by apache.

the class MasterRepairsDualAssignmentIT method test.

@Test
public void test() throws Exception {
    // make some tablets, spread 'em around
    Connector c = getConnector();
    ClientContext context = new ClientContext(c.getInstance(), new Credentials("root", new PasswordToken(ROOT_PASSWORD)), getClientConfig());
    String table = this.getUniqueNames(1)[0];
    c.securityOperations().grantTablePermission("root", MetadataTable.NAME, TablePermission.WRITE);
    c.securityOperations().grantTablePermission("root", RootTable.NAME, TablePermission.WRITE);
    c.tableOperations().create(table);
    SortedSet<Text> partitions = new TreeSet<>();
    for (String part : "a b c d e f g h i j k l m n o p q r s t u v w x y z".split(" ")) {
        partitions.add(new Text(part));
    }
    c.tableOperations().addSplits(table, partitions);
    // scan the metadata table and get the two table location states
    Set<TServerInstance> states = new HashSet<>();
    Set<TabletLocationState> oldLocations = new HashSet<>();
    MetaDataStateStore store = new MetaDataStateStore(context, null);
    while (states.size() < 2) {
        UtilWaitThread.sleep(250);
        oldLocations.clear();
        for (TabletLocationState tls : store) {
            if (tls.current != null) {
                states.add(tls.current);
                oldLocations.add(tls);
            }
        }
    }
    assertEquals(2, states.size());
    // Kill a tablet server... we don't care which one... wait for everything to be reassigned
    cluster.killProcess(ServerType.TABLET_SERVER, cluster.getProcesses().get(ServerType.TABLET_SERVER).iterator().next());
    Set<TServerInstance> replStates = new HashSet<>();
    // Find out which tablet server remains
    while (true) {
        UtilWaitThread.sleep(1000);
        states.clear();
        replStates.clear();
        boolean allAssigned = true;
        for (TabletLocationState tls : store) {
            if (tls != null && tls.current != null) {
                states.add(tls.current);
            } else if (tls != null && tls.extent.equals(new KeyExtent(ReplicationTable.ID, null, null))) {
                replStates.add(tls.current);
            } else {
                allAssigned = false;
            }
        }
        System.out.println(states + " size " + states.size() + " allAssigned " + allAssigned);
        if (states.size() != 2 && allAssigned)
            break;
    }
    assertEquals(1, replStates.size());
    assertEquals(1, states.size());
    // pick an assigned tablet and assign it to the old tablet
    TabletLocationState moved = null;
    for (TabletLocationState old : oldLocations) {
        if (!states.contains(old.current)) {
            moved = old;
        }
    }
    assertNotEquals(null, moved);
    // throw a mutation in as if we were the dying tablet
    BatchWriter bw = c.createBatchWriter(MetadataTable.NAME, new BatchWriterConfig());
    Mutation assignment = new Mutation(moved.extent.getMetadataEntry());
    moved.current.putLocation(assignment);
    bw.addMutation(assignment);
    bw.close();
    // wait for the master to fix the problem
    waitForCleanStore(store);
    // now jam up the metadata table
    bw = c.createBatchWriter(MetadataTable.NAME, new BatchWriterConfig());
    assignment = new Mutation(new KeyExtent(MetadataTable.ID, null, null).getMetadataEntry());
    moved.current.putLocation(assignment);
    bw.addMutation(assignment);
    bw.close();
    waitForCleanStore(new RootTabletStateStore(context, null));
}