/* This file is part of VoltDB.
* Copyright (C) 2008-2017 VoltDB Inc.
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
* IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
* OTHER DEALINGS IN THE SOFTWARE.
*/
package org.voltdb.iv2;
import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertTrue;
import static org.mockito.Matchers.any;
import static org.mockito.Matchers.anyLong;
import static org.mockito.Mockito.mock;
import static org.mockito.Mockito.reset;
import static org.mockito.Mockito.timeout;
import static org.mockito.Mockito.verify;
import static org.mockito.Mockito.when;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Set;
import org.apache.zookeeper_voltpatches.ZooKeeper;
import org.json_voltpatches.JSONObject;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import org.mockito.ArgumentCaptor;
import org.voltcore.messaging.BinaryPayloadMessage;
import org.voltcore.messaging.HostMessenger;
import org.voltcore.messaging.VoltMessage;
import org.voltcore.utils.CoreUtils;
import org.voltcore.zk.LeaderElector;
import org.voltcore.zk.ZKTestBase;
import org.voltdb.VoltDB;
import org.voltdb.VoltZK;
import com.google_voltpatches.common.collect.Lists;
import com.google_voltpatches.common.collect.Maps;
import com.google_voltpatches.common.collect.Sets;
public class TestCartographer extends ZKTestBase {
@Before
public void setUp() throws Exception
{
setUpZK(1);
}
@After
public void tearDown() throws Exception
{
tearDownZK();
}
@Test
public void testSortByValue()
{
Map<Integer, Integer> unsorted = new HashMap<Integer, Integer>();
unsorted.put(0, 5);
unsorted.put(1, 0);
unsorted.put(2, 3);
unsorted.put(3, 0);
unsorted.put(4, 2);
List<Integer> sorted = Cartographer.sortKeysByValue(unsorted);
assertEquals((Integer)1, sorted.get(0));
assertEquals((Integer)3, sorted.get(1));
assertEquals((Integer)4, sorted.get(2));
assertEquals((Integer)2, sorted.get(3));
assertEquals((Integer)0, sorted.get(4));
}
@Test
public void testReplicateLeastReplicated()
{
// Gin up our fake replication counts to leave the last partition needing two replicas
// 5 hosts, 3 sites/host, k=2 is 15 sites, 5 partitions.
// Need 2 hosts dead, so 6 missing replicas, last one needs to miss 2
Map<Integer, Integer> repsPerPart = new HashMap<Integer, Integer>();
repsPerPart.put(0, 2);
repsPerPart.put(1, 2);
repsPerPart.put(2, 2);
repsPerPart.put(3, 2);
repsPerPart.put(4, 1);
int kfactor = 2;
int sitesPerHost = 3;
List<Integer> firstRejoin = new ArrayList<Integer>();
Cartographer.computeReplacementPartitions(repsPerPart, kfactor, sitesPerHost, firstRejoin);
assertEquals(3, firstRejoin.size());
assertEquals((Integer)4, firstRejoin.get(0));
assertEquals((Integer)0, firstRejoin.get(1));
assertEquals((Integer)1, firstRejoin.get(2));
// add to each of the repsPerPart from the firstRejoin results
for (int partition : firstRejoin) {
repsPerPart.put(partition, repsPerPart.get(partition) + 1);
}
// now do it again and make sure we fully replicate
List<Integer> secondRejoin = new ArrayList<Integer>();
Cartographer.computeReplacementPartitions(repsPerPart, kfactor, sitesPerHost, secondRejoin);
assertEquals(3, secondRejoin.size());
assertEquals((Integer)2, secondRejoin.get(0));
assertEquals((Integer)3, secondRejoin.get(1));
assertEquals((Integer)4, secondRejoin.get(2));
}
@Test
public void testNoOverReplication()
{
// We'll set things up to only require one more replica and then offer up a host with more sites
// than we need. This particular case could never happen but should test the logic correctly.
// Also, test that we don't replicate the same partition twice on the same host again.
int kfactor = 2;
int sitesPerHost = 3;
Map<Integer, Integer> repsPerPart = new HashMap<Integer, Integer>();
repsPerPart.put(0, 3);
repsPerPart.put(1, 3);
repsPerPart.put(2, 3);
repsPerPart.put(3, 3);
repsPerPart.put(4, 1);
List<Integer> firstRejoin = new ArrayList<Integer>();
Cartographer.computeReplacementPartitions(repsPerPart, kfactor, sitesPerHost, firstRejoin);
assertEquals(1, firstRejoin.size());
assertEquals((Integer)4, firstRejoin.get(0));
}
@Test
public void testSPMasterChange() throws Exception
{
ZooKeeper zk = getClient(0);
VoltZK.createPersistentZKNodes(zk);
LeaderCache spwriter = new LeaderCache(zk, VoltZK.iv2masters);
HostMessenger hm = mock(HostMessenger.class);
when(hm.getZK()).thenReturn(m_messengers.get(0).getZK());
Cartographer dut = new Cartographer(hm, 0, false);
// Startup partitions
spwriter.start(true);
spwriter.put(0, 0l);
verify(hm, timeout(10000)).send(anyLong(), any(VoltMessage.class));
reset(hm);
spwriter.put(1, 1l);
verify(hm, timeout(10000)).send(anyLong(), any(VoltMessage.class));
reset(hm);
spwriter.put(2, 2l);
verify(hm, timeout(10000)).send(anyLong(), any(VoltMessage.class));
reset(hm);
// now change master for part 0
spwriter.put(0, 3l);
ArgumentCaptor<Long> hsIdCaptor = ArgumentCaptor.forClass(Long.class);
ArgumentCaptor<BinaryPayloadMessage> bpmCaptor = ArgumentCaptor.forClass(BinaryPayloadMessage.class);
verify(hm, timeout(10000)).send(hsIdCaptor.capture(), bpmCaptor.capture());
JSONObject jsObj = new JSONObject(new String(bpmCaptor.getValue().m_payload, "UTF-8"));
System.out.println("BPM: " + jsObj.toString());
final int partitionId = jsObj.getInt(Cartographer.JSON_PARTITION_ID);
final long initiatorHSId = jsObj.getLong(Cartographer.JSON_INITIATOR_HSID);
assertEquals(0, partitionId);
assertEquals(3, initiatorHSId);
spwriter.shutdown();
}
@Test
public void testMPIChange() throws Exception
{
ZooKeeper zk = getClient(0);
VoltZK.createPersistentZKNodes(zk);
LeaderCache mpwriter = new LeaderCache(zk, VoltZK.iv2mpi);
HostMessenger hm = mock(HostMessenger.class);
when(hm.getZK()).thenReturn(m_messengers.get(0).getZK());
Cartographer dut = new Cartographer(hm, 0, false);
mpwriter.start(true);
// initial master
mpwriter.put(MpInitiator.MP_INIT_PID, 0l);
verify(hm, timeout(10000)).send(anyLong(), any(VoltMessage.class));
reset(hm);
// Now change the master
mpwriter.put(MpInitiator.MP_INIT_PID, 3l);
ArgumentCaptor<Long> hsIdCaptor = ArgumentCaptor.forClass(Long.class);
ArgumentCaptor<BinaryPayloadMessage> bpmCaptor = ArgumentCaptor.forClass(BinaryPayloadMessage.class);
verify(hm, timeout(10000)).send(hsIdCaptor.capture(), bpmCaptor.capture());
JSONObject jsObj = new JSONObject(new String(bpmCaptor.getValue().m_payload, "UTF-8"));
final int partitionId = jsObj.getInt(Cartographer.JSON_PARTITION_ID);
final long initiatorHSId = jsObj.getLong(Cartographer.JSON_INITIATOR_HSID);
assertEquals(MpInitiator.MP_INIT_PID, partitionId);
assertEquals(3, initiatorHSId);
mpwriter.shutdown();
}
@Test
public void testRackAwareRejoin() throws Exception
{
ZooKeeper zk = getClient(0);
VoltZK.createPersistentZKNodes(zk);
HostMessenger hm = mock(HostMessenger.class);
when(hm.getZK()).thenReturn(m_messengers.get(0).getZK());
Cartographer dut = new Cartographer(hm, 0, false);
// In total there are 4 partitions, let's say there are two nodes missing
// and one is rejoining back
int kfactor = 1;
int sitesPerHost = 2;
int hostCount = 4;
int totalPartitions = (hostCount * sitesPerHost) / (kfactor + 1);
Map<Integer, String> hostGroups = Maps.newHashMap();
hostGroups.put(0, "rack1");
hostGroups.put(1, "rack1"); // this is rejoining node
hostGroups.put(2, "rack2");
// hostGroups.put(3, "rack2"); // this node is dead
Set<Integer> deadHosts = Sets.newHashSet();
deadHosts.add(1);
deadHosts.add(3);
int rejoiningHostId = 1;
// Depends on number of partitions this part can be slow.
int hostIdCounter = 0;
int[] siteIds = new int[hostCount];
for (int i = 0; i < hostCount; i++) {
siteIds[i] = 0;
}
// assign partitions
for (int pid = 0; pid < totalPartitions; pid++) {
LeaderElector.createRootIfNotExist(zk, LeaderElector.electionDirForPartition(VoltZK.leaders_initiators, pid));
Thread.sleep(500); // I'm evil
assertFalse(VoltDB.wasCrashCalled);
// add replica
for (int k = 0; k <= kfactor; k++) {
int hid = hostIdCounter++ % hostCount;
if (deadHosts.contains(hid)) continue;
long HSId = CoreUtils.getHSIdFromHostAndSite(hid, siteIds[hid]++);
LeaderElector.createParticipantNode(zk,
LeaderElector.electionDirForPartition(VoltZK.leaders_initiators, pid),
Long.toString(HSId++),
null);
}
}
/*
* Partition layout should be:
* H0: p0, p2
* H1: ?, ? (rejoining, expect p1, p3)
* H2: p1, p3
* H3: p0, p2 (dead)
*/
List<Integer> expectedPartitions = Lists.newArrayList();
expectedPartitions.add(1);
expectedPartitions.add(3);
List<Integer> partitionsToReplace = dut.getIv2PartitionsToReplace(kfactor, sitesPerHost, rejoiningHostId, hostGroups);
assertTrue(partitionsToReplace.size() == sitesPerHost);
for (Integer p : partitionsToReplace) {
assertTrue(expectedPartitions.contains(p));
}
}
}