Search in sources :

Example 11 with TabletMigration

use of org.apache.accumulo.core.spi.balancer.data.TabletMigration in project accumulo by apache.

the class HostRegexTableLoadBalancer method balance.

@Override
public long balance(BalanceParameters params) {
    long minBalanceTime = 20_000;
    // Iterate over the tables and balance each of them
    Map<String, TableId> tableIdMap = environment.getTableIdMap();
    Map<TableId, String> tableIdToTableName = tableIdMap.entrySet().stream().collect(Collectors.toMap(Map.Entry::getValue, Map.Entry::getKey));
    tableIdToTableName.keySet().forEach(this::checkTableConfig);
    long now = System.currentTimeMillis();
    HrtlbConf myConf = hrtlbConf.get();
    SortedMap<TabletServerId, TServerStatus> current = params.currentStatus();
    Set<TabletId> migrations = params.currentMigrations();
    List<TabletMigration> migrationsOut = params.migrationsOut();
    Map<String, SortedMap<TabletServerId, TServerStatus>> currentGrouped = splitCurrentByRegex(params.currentStatus());
    if ((now - this.lastOOBCheck) > myConf.oobCheckMillis) {
        try {
            // Check to see if a tablet is assigned outside the bounds of the pool. If so, migrate it.
            for (String table : tableIdMap.keySet()) {
                LOG.debug("Checking for out of bounds tablets for table {}", table);
                String tablePoolName = getPoolNameForTable(table);
                for (Entry<TabletServerId, TServerStatus> e : current.entrySet()) {
                    // pool names are the same as table names, except in the DEFAULT case.
                    // If this table is assigned to a pool for this host, then move on.
                    List<String> hostPools = getPoolNamesForHost(e.getKey());
                    if (hostPools.contains(tablePoolName)) {
                        continue;
                    }
                    TableId tid = tableIdMap.get(table);
                    if (tid == null) {
                        LOG.warn("Unable to check for out of bounds tablets for table {}," + " it may have been deleted or renamed.", table);
                        continue;
                    }
                    try {
                        List<TabletStatistics> outOfBoundsTablets = getOnlineTabletsForTable(e.getKey(), tid);
                        if (outOfBoundsTablets == null) {
                            continue;
                        }
                        for (TabletStatistics ts : outOfBoundsTablets) {
                            if (migrations.contains(ts.getTabletId())) {
                                LOG.debug("Migration for out of bounds tablet {} has already been requested", ts.getTabletId());
                                continue;
                            }
                            String poolName = getPoolNameForTable(table);
                            SortedMap<TabletServerId, TServerStatus> currentView = currentGrouped.get(poolName);
                            if (currentView != null) {
                                int skip = random.nextInt(currentView.size());
                                Iterator<TabletServerId> iter = currentView.keySet().iterator();
                                for (int i = 0; i < skip; i++) {
                                    iter.next();
                                }
                                TabletServerId nextTS = iter.next();
                                LOG.info("Tablet {} is currently outside the bounds of the" + " regex, migrating from {} to {}", ts.getTabletId(), e.getKey(), nextTS);
                                migrationsOut.add(new TabletMigration(ts.getTabletId(), e.getKey(), nextTS));
                                if (migrationsOut.size() >= myConf.maxTServerMigrations) {
                                    break;
                                }
                            } else {
                                LOG.warn("No tablet servers online for pool {}, unable to" + " migrate out of bounds tablets", poolName);
                            }
                        }
                    } catch (AccumuloException | AccumuloSecurityException e1) {
                        LOG.error("Error in OOB check getting tablets for table {} from server {} {}", tid, e.getKey().getHost(), e);
                    }
                }
            }
        } finally {
            // this could have taken a while...get a new time
            this.lastOOBCheck = System.currentTimeMillis();
        }
    }
    if (!migrationsOut.isEmpty()) {
        LOG.warn("Not balancing tables due to moving {} out of bounds tablets", migrationsOut.size());
        LOG.info("Migrating out of bounds tablets: {}", migrationsOut);
        return minBalanceTime;
    }
    if (migrations != null && !migrations.isEmpty()) {
        if (migrations.size() >= myConf.maxOutstandingMigrations) {
            LOG.warn("Not balancing tables due to {} outstanding migrations", migrations.size());
            if (LOG.isTraceEnabled()) {
                LOG.trace("Sample up to 10 outstanding migrations: {}", Iterables.limit(migrations, 10));
            }
            return minBalanceTime;
        }
        LOG.debug("Current outstanding migrations of {} being applied", migrations.size());
        if (LOG.isTraceEnabled()) {
            LOG.trace("Sample up to 10 outstanding migrations: {}", Iterables.limit(migrations, 10));
        }
        migrationsFromLastPass.keySet().retainAll(migrations);
        SortedMap<TabletServerId, TServerStatusImpl> currentCopy = new TreeMap<>();
        current.forEach((tid, status) -> currentCopy.put(tid, (TServerStatusImpl) status));
        Multimap<TabletServerId, String> serverTableIdCopied = HashMultimap.create();
        for (TabletMigration migration : migrationsFromLastPass.values()) {
            TableStatisticsImpl fromInfo = getTableInfo(currentCopy, serverTableIdCopied, migration.getTablet().getTable().canonical(), migration.getOldTabletServer());
            if (fromInfo != null) {
                fromInfo.setOnlineTabletCount(fromInfo.getOnlineTabletCount() - 1);
            }
            TableStatisticsImpl toInfo = getTableInfo(currentCopy, serverTableIdCopied, migration.getTablet().getTable().canonical(), migration.getNewTabletServer());
            if (toInfo != null) {
                toInfo.setOnlineTabletCount(toInfo.getOnlineTabletCount() + 1);
            }
        }
        migrations = EMPTY_MIGRATIONS;
    } else {
        migrationsFromLastPass.clear();
    }
    for (TableId tableId : tableIdMap.values()) {
        String tableName = tableIdToTableName.get(tableId);
        String regexTableName = getPoolNameForTable(tableName);
        SortedMap<TabletServerId, TServerStatus> currentView = currentGrouped.get(regexTableName);
        if (currentView == null) {
            LOG.warn("Skipping balance for table {} as no tablet servers are online.", tableName);
            continue;
        }
        ArrayList<TabletMigration> newMigrations = new ArrayList<>();
        getBalancerForTable(tableId).balance(new BalanceParamsImpl(currentView, migrations, newMigrations));
        if (newMigrations.isEmpty()) {
            tableToTimeSinceNoMigrations.remove(tableId);
        } else if (tableToTimeSinceNoMigrations.containsKey(tableId)) {
            if ((now - tableToTimeSinceNoMigrations.get(tableId)) > HOURS.toMillis(1)) {
                LOG.warn("We have been consistently producing migrations for {}: {}", tableName, Iterables.limit(newMigrations, 10));
            }
        } else {
            tableToTimeSinceNoMigrations.put(tableId, now);
        }
        migrationsOut.addAll(newMigrations);
        if (migrationsOut.size() >= myConf.maxTServerMigrations) {
            break;
        }
    }
    for (TabletMigration migration : migrationsOut) {
        migrationsFromLastPass.put(migration.getTablet(), migration);
    }
    LOG.info("Migrating tablets for balance: {}", migrationsOut);
    return minBalanceTime;
}
Also used : TableId(org.apache.accumulo.core.data.TableId) ArrayList(java.util.ArrayList) TServerStatusImpl(org.apache.accumulo.core.manager.balancer.TServerStatusImpl) TabletServerId(org.apache.accumulo.core.spi.balancer.data.TabletServerId) TabletStatistics(org.apache.accumulo.core.spi.balancer.data.TabletStatistics) AccumuloSecurityException(org.apache.accumulo.core.client.AccumuloSecurityException) AccumuloException(org.apache.accumulo.core.client.AccumuloException) BalanceParamsImpl(org.apache.accumulo.core.manager.balancer.BalanceParamsImpl) TabletMigration(org.apache.accumulo.core.spi.balancer.data.TabletMigration) TServerStatus(org.apache.accumulo.core.spi.balancer.data.TServerStatus) TreeMap(java.util.TreeMap) TableStatisticsImpl(org.apache.accumulo.core.manager.balancer.TableStatisticsImpl) SortedMap(java.util.SortedMap) TabletId(org.apache.accumulo.core.data.TabletId) HashMap(java.util.HashMap) Map(java.util.Map) TreeMap(java.util.TreeMap) SortedMap(java.util.SortedMap)

Example 12 with TabletMigration

use of org.apache.accumulo.core.spi.balancer.data.TabletMigration in project accumulo by apache.

the class TableLoadBalancer method balance.

@Override
public long balance(BalanceParameters params) {
    long minBalanceTime = 5_000;
    // Iterate over the tables and balance each of them
    for (TableId tableId : environment.getTableIdMap().values()) {
        ArrayList<TabletMigration> newMigrations = new ArrayList<>();
        long tableBalanceTime = getBalancerForTable(tableId).balance(new BalanceParamsImpl(params.currentStatus(), params.currentMigrations(), newMigrations));
        if (tableBalanceTime < minBalanceTime)
            minBalanceTime = tableBalanceTime;
        params.migrationsOut().addAll(newMigrations);
    }
    return minBalanceTime;
}
Also used : TableId(org.apache.accumulo.core.data.TableId) BalanceParamsImpl(org.apache.accumulo.core.manager.balancer.BalanceParamsImpl) TabletMigration(org.apache.accumulo.core.spi.balancer.data.TabletMigration) ArrayList(java.util.ArrayList)

Example 13 with TabletMigration

use of org.apache.accumulo.core.spi.balancer.data.TabletMigration in project accumulo by apache.

the class HostRegexTableLoadBalancerTest method testOutOfBoundsTablets.

@Test
public void testOutOfBoundsTablets() {
    init(DEFAULT_TABLE_PROPERTIES);
    // Wait to trigger the out of bounds check which will call our version of
    // getOnlineTabletsForTable
    UtilWaitThread.sleep(11000);
    Set<TabletId> migrations = new HashSet<>();
    List<TabletMigration> migrationsOut = new ArrayList<>();
    this.balance(new BalanceParamsImpl(createCurrent(15), migrations, migrationsOut));
    assertEquals(2, migrationsOut.size());
}
Also used : BalanceParamsImpl(org.apache.accumulo.core.manager.balancer.BalanceParamsImpl) TabletMigration(org.apache.accumulo.core.spi.balancer.data.TabletMigration) ArrayList(java.util.ArrayList) TabletId(org.apache.accumulo.core.data.TabletId) HashSet(java.util.HashSet) Test(org.junit.jupiter.api.Test)

Example 14 with TabletMigration

use of org.apache.accumulo.core.spi.balancer.data.TabletMigration in project accumulo by apache.

the class SimpleLoadBalancerTest method testUnevenAssignment2.

@Test
public void testUnevenAssignment2() {
    // make 26 servers
    for (char c : "abcdefghijklmnopqrstuvwxyz".toCharArray()) {
        TabletServerId tsid = new TabletServerIdImpl("127.0.0.1", c, Character.toString(c));
        FakeTServer fakeTServer = new FakeTServer();
        servers.put(tsid, fakeTServer);
    }
    // put 60 tablets on 25 of them
    List<Entry<TabletServerId, FakeTServer>> shortList = new ArrayList<>(servers.entrySet());
    Entry<TabletServerId, FakeTServer> shortServer = shortList.remove(0);
    int c = 0;
    for (int i = 0; i < 60; i++) {
        for (Entry<TabletServerId, FakeTServer> entry : shortList) {
            entry.getValue().tablets.add(makeTablet("t" + c, null, null));
        }
    }
    // put 10 on the that short server:
    for (int i = 0; i < 10; i++) {
        shortServer.getValue().tablets.add(makeTablet("s" + i, null, null));
    }
    TestSimpleLoadBalancer balancer = new TestSimpleLoadBalancer();
    Set<TabletId> migrations = Collections.emptySet();
    int moved = 0;
    // balance until we can't balance no more!
    while (true) {
        List<TabletMigration> migrationsOut = new ArrayList<>();
        balancer.balance(new BalanceParamsImpl(getAssignments(servers), migrations, migrationsOut));
        if (migrationsOut.isEmpty())
            break;
        for (TabletMigration migration : migrationsOut) {
            if (servers.get(migration.getOldTabletServer()).tablets.remove(migration.getTablet()))
                moved++;
            last.remove(migration.getTablet());
            servers.get(migration.getNewTabletServer()).tablets.add(migration.getTablet());
            last.put(migration.getTablet(), migration.getNewTabletServer());
        }
    }
    // average is 58, with 2 at 59: we need 48 more moved to the short server
    assertEquals(48, moved);
}
Also used : BalanceParamsImpl(org.apache.accumulo.core.manager.balancer.BalanceParamsImpl) TabletMigration(org.apache.accumulo.core.spi.balancer.data.TabletMigration) ArrayList(java.util.ArrayList) TabletServerIdImpl(org.apache.accumulo.core.manager.balancer.TabletServerIdImpl) Entry(java.util.Map.Entry) TabletServerId(org.apache.accumulo.core.spi.balancer.data.TabletServerId) TabletId(org.apache.accumulo.core.data.TabletId) Test(org.junit.jupiter.api.Test)

Example 15 with TabletMigration

use of org.apache.accumulo.core.spi.balancer.data.TabletMigration in project accumulo by apache.

the class SimpleLoadBalancerTest method testUnevenAssignment.

@Test
public void testUnevenAssignment() {
    for (char c : "abcdefghijklmnopqrstuvwxyz".toCharArray()) {
        String cString = Character.toString(c);
        TabletServerId tsid = new TabletServerIdImpl("127.0.0.1", c, cString);
        FakeTServer fakeTServer = new FakeTServer();
        servers.put(tsid, fakeTServer);
        fakeTServer.tablets.add(makeTablet(cString, null, null));
    }
    // Put more tablets on one server, but not more than the number of servers
    Entry<TabletServerId, FakeTServer> first = servers.entrySet().iterator().next();
    first.getValue().tablets.add(makeTablet("newTable", "a", null));
    first.getValue().tablets.add(makeTablet("newTable", "b", "a"));
    first.getValue().tablets.add(makeTablet("newTable", "c", "b"));
    first.getValue().tablets.add(makeTablet("newTable", "d", "c"));
    first.getValue().tablets.add(makeTablet("newTable", "e", "d"));
    first.getValue().tablets.add(makeTablet("newTable", "f", "e"));
    first.getValue().tablets.add(makeTablet("newTable", "g", "f"));
    first.getValue().tablets.add(makeTablet("newTable", "h", "g"));
    first.getValue().tablets.add(makeTablet("newTable", "i", null));
    TestSimpleLoadBalancer balancer = new TestSimpleLoadBalancer();
    Set<TabletId> migrations = Collections.emptySet();
    int moved = 0;
    // balance until we can't balance no more!
    while (true) {
        List<TabletMigration> migrationsOut = new ArrayList<>();
        balancer.balance(new BalanceParamsImpl(getAssignments(servers), migrations, migrationsOut));
        if (migrationsOut.isEmpty())
            break;
        for (TabletMigration migration : migrationsOut) {
            if (servers.get(migration.getOldTabletServer()).tablets.remove(migration.getTablet()))
                moved++;
            servers.get(migration.getNewTabletServer()).tablets.add(migration.getTablet());
        }
    }
    assertEquals(8, moved);
}
Also used : BalanceParamsImpl(org.apache.accumulo.core.manager.balancer.BalanceParamsImpl) TabletMigration(org.apache.accumulo.core.spi.balancer.data.TabletMigration) ArrayList(java.util.ArrayList) TabletServerIdImpl(org.apache.accumulo.core.manager.balancer.TabletServerIdImpl) TabletServerId(org.apache.accumulo.core.spi.balancer.data.TabletServerId) TabletId(org.apache.accumulo.core.data.TabletId) Test(org.junit.jupiter.api.Test)

Aggregations

TabletMigration (org.apache.accumulo.core.spi.balancer.data.TabletMigration)15 ArrayList (java.util.ArrayList)13 TabletId (org.apache.accumulo.core.data.TabletId)13 BalanceParamsImpl (org.apache.accumulo.core.manager.balancer.BalanceParamsImpl)10 TabletServerId (org.apache.accumulo.core.spi.balancer.data.TabletServerId)10 Test (org.junit.jupiter.api.Test)8 TableId (org.apache.accumulo.core.data.TableId)6 HashMap (java.util.HashMap)5 TabletServerIdImpl (org.apache.accumulo.core.manager.balancer.TabletServerIdImpl)5 HashSet (java.util.HashSet)4 TServerStatus (org.apache.accumulo.core.spi.balancer.data.TServerStatus)4 TabletStatistics (org.apache.accumulo.core.spi.balancer.data.TabletStatistics)4 TreeMap (java.util.TreeMap)3 AccumuloException (org.apache.accumulo.core.client.AccumuloException)3 AccumuloSecurityException (org.apache.accumulo.core.client.AccumuloSecurityException)3 Map (java.util.Map)2 SortedMap (java.util.SortedMap)2 ConfigurationCopy (org.apache.accumulo.core.conf.ConfigurationCopy)2 TServerStatusImpl (org.apache.accumulo.core.manager.balancer.TServerStatusImpl)2 ConfigurationImpl (org.apache.accumulo.core.util.ConfigurationImpl)2