Jackie-Jiang commented on code in PR #11568: URL: https://github.com/apache/pinot/pull/11568#discussion_r1326702312
########## pinot-integration-tests/src/test/java/org/apache/pinot/integration/tests/UpsertTableRebalanceIntegrationTest.java: ########## @@ -0,0 +1,348 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.pinot.integration.tests; + +import java.io.File; +import java.io.IOException; +import java.util.HashMap; +import java.util.HashSet; +import java.util.List; +import java.util.Map; +import java.util.Set; +import org.apache.commons.configuration.BaseConfiguration; +import org.apache.commons.configuration.Configuration; +import org.apache.helix.model.IdealState; +import org.apache.pinot.common.utils.LLCSegmentName; +import org.apache.pinot.common.utils.helix.HelixHelper; +import org.apache.pinot.controller.helix.core.PinotHelixResourceManager; +import org.apache.pinot.controller.helix.core.rebalance.RebalanceResult; +import org.apache.pinot.controller.helix.core.rebalance.TableRebalancer; +import org.apache.pinot.server.starter.helix.BaseServerStarter; +import org.apache.pinot.spi.config.table.TableConfig; +import org.apache.pinot.spi.config.table.TableType; +import org.apache.pinot.spi.data.Schema; +import org.apache.pinot.spi.utils.CommonConstants; +import org.apache.pinot.spi.utils.RebalanceConfigConstants; +import org.apache.pinot.spi.utils.builder.TableNameBuilder; +import org.apache.pinot.util.TestUtils; +import org.testng.annotations.AfterClass; +import org.testng.annotations.AfterMethod; +import org.testng.annotations.BeforeClass; +import org.testng.annotations.Test; + +import static org.testng.Assert.assertEquals; + +public class UpsertTableRebalanceIntegrationTest extends BaseClusterIntegrationTest { Review Comment: Partial upsert should cover all the functionality of the regular upsert table. Since each integration test takes a long time to setup, I'd suggest just keeping the partial upsert test. Ideally we want to enhance the existing upsert integration test to also test the rebalance functionality ########## pinot-integration-tests/src/test/java/org/apache/pinot/integration/tests/PartialUpsertTableRebalanceIntegrationTest.java: ########## @@ -0,0 +1,371 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.pinot.integration.tests; + +import java.io.File; +import java.io.IOException; +import java.util.HashMap; +import java.util.HashSet; +import java.util.List; +import java.util.Map; +import java.util.Set; +import org.apache.commons.configuration.BaseConfiguration; +import org.apache.commons.configuration.Configuration; +import org.apache.helix.model.IdealState; +import org.apache.pinot.client.ResultSetGroup; +import org.apache.pinot.common.utils.LLCSegmentName; +import org.apache.pinot.common.utils.helix.HelixHelper; +import org.apache.pinot.controller.helix.core.PinotHelixResourceManager; +import org.apache.pinot.controller.helix.core.rebalance.RebalanceResult; +import org.apache.pinot.controller.helix.core.rebalance.TableRebalancer; +import org.apache.pinot.server.starter.helix.BaseServerStarter; +import org.apache.pinot.spi.config.table.TableConfig; +import org.apache.pinot.spi.config.table.TableType; +import org.apache.pinot.spi.data.Schema; +import org.apache.pinot.spi.utils.CommonConstants; +import org.apache.pinot.spi.utils.RebalanceConfigConstants; +import org.apache.pinot.spi.utils.builder.TableNameBuilder; +import org.apache.pinot.util.TestUtils; +import org.testng.annotations.AfterClass; +import org.testng.annotations.AfterMethod; +import org.testng.annotations.BeforeClass; +import org.testng.annotations.Test; + +import static org.testng.Assert.assertEquals; + + +public class PartialUpsertTableRebalanceIntegrationTest extends BaseClusterIntegrationTest { + private static final int NUM_SERVERS = 1; + private static final String PRIMARY_KEY_COL = "clientId"; + private static final String REALTIME_TABLE_NAME = TableNameBuilder.REALTIME.tableNameWithType(DEFAULT_TABLE_NAME); + + // Segment 1 contains records of pk value 100000 (partition 0) + private static final String UPLOADED_SEGMENT_1 = "mytable_10027_19736_0 %"; + // Segment 2 contains records of pk value 100001 (partition 1) + private static final String UPLOADED_SEGMENT_2 = "mytable_10072_19919_1 %"; + // Segment 3 contains records of pk value 100002 (partition 1) + private static final String UPLOADED_SEGMENT_3 = "mytable_10158_19938_2 %"; + + private PinotHelixResourceManager _resourceManager; + private TableRebalancer _tableRebalancer; + private static List<File> _avroFiles; + private TableConfig _tableConfig; + private Schema _schema; + + @BeforeClass + public void setUp() + throws Exception { + TestUtils.ensureDirectoriesExistAndEmpty(_tempDir, _segmentDir, _tarDir); + + startZk(); + startController(); + startBroker(); + startServers(NUM_SERVERS); + + // Start Kafka and push data into Kafka + startKafka(); + + _resourceManager = getControllerStarter().getHelixResourceManager(); + _tableRebalancer = new TableRebalancer(_resourceManager.getHelixZkManager(), null); + + createSchemaAndTable(); + } + + @Test + public void testRebalance() + throws Exception { + populateTables(); + + verifyIdealState(5, NUM_SERVERS); + + // setup the rebalance config + Configuration rebalanceConfig = new BaseConfiguration(); + rebalanceConfig.addProperty(RebalanceConfigConstants.DRY_RUN, false); + rebalanceConfig.addProperty(RebalanceConfigConstants.MIN_REPLICAS_TO_KEEP_UP_FOR_NO_DOWNTIME, 0); + rebalanceConfig.addProperty(RebalanceConfigConstants.INCLUDE_CONSUMING, true); + + // Add a new server + BaseServerStarter serverStarter1 = startOneServer(1234); + + // Now we trigger a rebalance operation + TableConfig tableConfig = _resourceManager.getTableConfig(REALTIME_TABLE_NAME); + RebalanceResult rebalanceResult = _tableRebalancer.rebalance(tableConfig, rebalanceConfig); + + // Check the number of replicas after rebalancing + int finalReplicas = _resourceManager.getServerInstancesForTable(getTableName(), TableType.REALTIME).size(); + + // Check that a replica has been added + assertEquals(finalReplicas, NUM_SERVERS + 1, "Rebalancing didn't correctly add the new server"); + + verifySegmentAssignment(rebalanceResult.getSegmentAssignment(), 5, finalReplicas); + + // Add a new server + BaseServerStarter serverStarter2 = startOneServer(4567); + rebalanceResult = _tableRebalancer.rebalance(tableConfig, rebalanceConfig); + + // Check the number of replicas after rebalancing + finalReplicas = _resourceManager.getServerInstancesForTable(getTableName(), TableType.REALTIME).size(); + + // Check that a replica has been added + assertEquals(finalReplicas, NUM_SERVERS + 2, "Rebalancing didn't correctly add the new server"); + + waitForAllDocsLoaded(600_000L); + + // number of instances assigned can't be more than number of partitions for rf = 1 + verifySegmentAssignment(rebalanceResult.getSegmentAssignment(), 5, getNumKafkaPartitions()); + + _resourceManager.updateInstanceTags(serverStarter1.getInstanceId(), "", false); + _resourceManager.updateInstanceTags(serverStarter2.getInstanceId(), "", false); + + rebalanceConfig.addProperty(RebalanceConfigConstants.REASSIGN_INSTANCES, true); + rebalanceConfig.addProperty(RebalanceConfigConstants.DOWNTIME, true); + + rebalanceResult = _tableRebalancer.rebalance(tableConfig, rebalanceConfig); + + verifySegmentAssignment(rebalanceResult.getSegmentAssignment(), 5, NUM_SERVERS); + + waitForAllDocsLoaded(600_000L); + + _resourceManager.disableInstance(serverStarter1.getInstanceId()); + _resourceManager.disableInstance(serverStarter2.getInstanceId()); + + _resourceManager.dropInstance(serverStarter1.getInstanceId()); + _resourceManager.dropInstance(serverStarter2.getInstanceId()); + + serverStarter1.stop(); + serverStarter2.stop(); + } + + @Test + public void testReload() + throws Exception { + pushAvroIntoKafka(_avroFiles); + reloadRealtimeTable(getTableName()); + waitForAllDocsLoaded(600_000L, 300); + verifyIdealState(2, NUM_SERVERS); + } + + @AfterMethod + public void afterMethod() + throws Exception { + String realtimeTableName = TableNameBuilder.REALTIME.tableNameWithType(getTableName()); + + // Test dropping all segments one by one + List<String> segments = listSegments(realtimeTableName); + for (String segment : segments) { + dropSegment(realtimeTableName, segment); + } + // NOTE: There is a delay to remove the segment from property store + TestUtils.waitForCondition((aVoid) -> { + try { + return listSegments(realtimeTableName).isEmpty(); + } catch (IOException e) { + throw new RuntimeException(e); + } + }, 60_000L, "Failed to drop the segments"); + + stopServer(); + stopKafka(); // to clean up the topic + startServers(NUM_SERVERS); + startKafka(); + } + + protected void verifySegmentAssignment(Map<String, Map<String, String>> segmentAssignment, int numSegmentsExpected, + int numInstancesExpected) { + assertEquals(segmentAssignment.size(), numSegmentsExpected); + + int maxSequenceNumber = 0; + for (Map.Entry<String, Map<String, String>> entry : segmentAssignment.entrySet()) { + String segmentName = entry.getKey(); + if (LLCSegmentName.isLowLevelConsumerSegmentName(segmentName)) { + LLCSegmentName llcSegmentName = new LLCSegmentName(segmentName); + maxSequenceNumber = Math.max(maxSequenceNumber, llcSegmentName.getSequenceNumber()); + } + } + + Map<Integer, String> serverForPartition = new HashMap<>(); + Set<String> uniqueServers = new HashSet<>(); + for (Map.Entry<String, Map<String, String>> entry : segmentAssignment.entrySet()) { + String segmentName = entry.getKey(); + Map<String, String> instanceStateMap = entry.getValue(); + + // Verify that all segments have the correct state + assertEquals(instanceStateMap.size(), 1); + Map.Entry<String, String> instanceIdAndState = instanceStateMap.entrySet().iterator().next(); + String state = instanceIdAndState.getValue(); + if (LLCSegmentName.isLowLevelConsumerSegmentName(segmentName)) { + LLCSegmentName llcSegmentName = new LLCSegmentName(segmentName); + if (llcSegmentName.getSequenceNumber() < maxSequenceNumber) { + assertEquals(state, CommonConstants.Helix.StateModel.SegmentStateModel.ONLINE); + } else { + assertEquals(state, CommonConstants.Helix.StateModel.SegmentStateModel.CONSUMING); + } + } else { + assertEquals(state, CommonConstants.Helix.StateModel.SegmentStateModel.ONLINE); + } + + // Verify that all segments of the same partition are mapped to the same server + String instanceId = instanceIdAndState.getKey(); + int partitionId = getSegmentPartitionId(segmentName); + uniqueServers.add(instanceId); + serverForPartition.compute(partitionId, (key, value) -> { + if (value == null) { + return instanceId; + } else { + assertEquals(instanceId, value); + return value; + } + }); + } + + assertEquals(uniqueServers.size(), numInstancesExpected); + } + + protected void verifyIdealState(int numSegmentsExpected, int numInstancesExpected) { + IdealState idealState = HelixHelper.getTableIdealState(_helixManager, REALTIME_TABLE_NAME); + verifySegmentAssignment(idealState.getRecord().getMapFields(), numSegmentsExpected, numInstancesExpected); + } + + protected void populateTables() + throws Exception { + // Create and upload segments + ClusterIntegrationTestUtils.buildSegmentsFromAvro(_avroFiles, _tableConfig, _schema, 0, _segmentDir, _tarDir); + uploadSegments(getTableName(), TableType.REALTIME, _tarDir); + + pushAvroIntoKafka(_avroFiles); + // Wait for all documents loaded + waitForAllDocsLoaded(600_000L); + } + + protected void createSchemaAndTable() + throws Exception { + // Unpack the Avro files + _avroFiles = unpackAvroData(_tempDir); + + // Create and upload schema and table config + _schema = createSchema(); + addSchema(_schema); + _tableConfig = createUpsertTableConfig(_avroFiles.get(0), PRIMARY_KEY_COL, null, getNumKafkaPartitions()); + _tableConfig.getValidationConfig().setDeletedSegmentsRetentionPeriod(null); + + addTableConfig(_tableConfig); + } + + @AfterClass + public void tearDown() { + stopServer(); + stopBroker(); + stopController(); + stopZk(); + } + + @Override + protected String getSchemaFileName() { + return "upsert_upload_segment_test.schema"; + } + + @Override + protected String getSchemaName() { + return "upsertSchema"; + } + + @Override + protected String getAvroTarFileName() { + return "upsert_upload_segment_test.tar.gz"; + } + + @Override + protected String getPartitionColumn() { + return PRIMARY_KEY_COL; + } + + @Override + protected long getCountStarResult() { + // Three distinct records are expected with pk values of 100000, 100001, 100002 + return 3; + } + + protected void waitForAllDocsLoaded(long timeoutMs, long expectedCount) + throws Exception { + TestUtils.waitForCondition(aVoid -> { + try { + return getCurrentCountStarResultWithoutUpsert() == expectedCount; + } catch (Exception e) { + return null; + } + }, 100L, timeoutMs, "Failed to load all documents"); + } + + private static int getSegmentPartitionId(String segmentName) { + switch (segmentName) { + case UPLOADED_SEGMENT_1: + return 0; + case UPLOADED_SEGMENT_2: + case UPLOADED_SEGMENT_3: + return 1; + default: + return new LLCSegmentName(segmentName).getPartitionGroupId(); + } + } + + @Override + protected void waitForAllDocsLoaded(long timeoutMs) + throws Exception { + Thread.sleep(5000L); //TODO: Wait for some time for rebalance to kickoff Review Comment: We want to avoid sleeping a fixed amount of time. For fast machine, it can waste time; for slow machine, it can cause flakiness. Any specific reason why waiting for condition itself is not enough? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
