swagle commented on a change in pull request #819: HDDS-1501 : Create a Recon task interface to update internal DB on updates from OM. URL: https://github.com/apache/hadoop/pull/819#discussion_r284480734
########## File path: hadoop-ozone/ozone-recon/src/main/java/org/apache/hadoop/ozone/recon/tasks/ReconTaskControllerImpl.java ########## @@ -0,0 +1,199 @@ +/** + * 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 + * <p> + * http://www.apache.org/licenses/LICENSE-2.0 + * <p> + * 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.hadoop.ozone.recon.tasks; + +import static org.apache.hadoop.ozone.recon.ReconServerConfigKeys.OZONE_RECON_TASK_THREAD_COUNT_DEFAULT; +import static org.apache.hadoop.ozone.recon.ReconServerConfigKeys.OZONE_RECON_TASK_THREAD_COUNT_KEY; + +import java.util.ArrayList; +import java.util.Collection; +import java.util.HashMap; +import java.util.List; +import java.util.Map; +import java.util.concurrent.Callable; +import java.util.concurrent.ExecutionException; +import java.util.concurrent.ExecutorService; +import java.util.concurrent.Executors; +import java.util.concurrent.Future; +import java.util.concurrent.Semaphore; +import java.util.concurrent.atomic.AtomicInteger; + +import org.apache.commons.lang3.tuple.Pair; +import org.apache.hadoop.hdds.conf.OzoneConfiguration; +import org.apache.hadoop.ozone.recon.recovery.ReconOMMetadataManager; +import org.hadoop.ozone.recon.schema.tables.daos.ReconTaskStatusDao; +import org.hadoop.ozone.recon.schema.tables.pojos.ReconTaskStatus; +import org.jooq.Configuration; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import com.google.inject.Inject; + +/** + * Implementation of ReconTaskController. + */ +public class ReconTaskControllerImpl implements ReconTaskController { + + private static final Logger LOG = + LoggerFactory.getLogger(ReconTaskControllerImpl.class); + + private Map<String, ReconDBUpdateTask> reconDBUpdateTasks; + private ExecutorService executorService; + private int threadCount = 1; + private final Semaphore taskSemaphore = new Semaphore(1); + private final ReconOMMetadataManager omMetadataManager; + private Map<String, AtomicInteger> taskFailureCounter = new HashMap<>(); + private static final int TASK_FAILURE_THRESHOLD = 2; + private ReconTaskStatusDao reconTaskStatusDao; + + @Inject + public ReconTaskControllerImpl(OzoneConfiguration configuration, + ReconOMMetadataManager omMetadataManager, + Configuration sqlConfiguration) { + this.omMetadataManager = omMetadataManager; + reconDBUpdateTasks = new HashMap<>(); + threadCount = configuration.getInt(OZONE_RECON_TASK_THREAD_COUNT_KEY, + OZONE_RECON_TASK_THREAD_COUNT_DEFAULT); + executorService = Executors.newFixedThreadPool(threadCount); + reconTaskStatusDao = new ReconTaskStatusDao(sqlConfiguration); + } + + @Override + public void registerTask(ReconDBUpdateTask task) { + String taskName = task.getTaskName(); + LOG.info("Registered task " + taskName + " with controller."); + + // Store task in Task Map. + reconDBUpdateTasks.put(taskName, task); + // Store Task in Task failure tracker. + taskFailureCounter.put(taskName, new AtomicInteger(0)); + // Create DB record for the task. + ReconTaskStatus reconTaskStatusRecord = new ReconTaskStatus(taskName, + 0L, 0L); + reconTaskStatusDao.insert(reconTaskStatusRecord); + } + + /** + * For every registered task, we try process step twice and then reprocess + * once (if process failed twice) to absorb the events. If a task has failed + * reprocess call more than 2 times across events, it is unregistered + * (blacklisted). + * @param events set of events + * @throws InterruptedException + */ + @Override + public void consumeOMEvents(OMUpdateEventBatch events) + throws InterruptedException { + + + taskSemaphore.acquire(); + + try { + Collection<Callable<Pair>> tasks = new ArrayList<>(); + for (Map.Entry<String, ReconDBUpdateTask> taskEntry : + reconDBUpdateTasks.entrySet()) { + ReconDBUpdateTask task = taskEntry.getValue(); + tasks.add(() -> task.process(events)); + } + + List<String> failedTasks = new ArrayList<>(); + List<Future<Pair>> results = executorService.invokeAll(tasks); + for (Future<Pair> f : results) { + String taskName = f.get().getLeft().toString(); + if (!(Boolean)f.get().getRight()) { + failedTasks.add(f.get().getLeft().toString()); + } else { + taskFailureCounter.get(taskName).set(0); + storeLastCompletedTransaction(taskName, events.getLastEventInfo()); + } + + } + + //Retry + List<String> retryFailedTasks = new ArrayList<>(); + if (!failedTasks.isEmpty()) { + tasks.clear(); + for (String taskName : failedTasks) { + ReconDBUpdateTask task = reconDBUpdateTasks.get(taskName); + tasks.add(() -> task.process(events)); + } + results = executorService.invokeAll(tasks); + for (Future<Pair> f : results) { + String taskName = f.get().getLeft().toString(); + if (!(Boolean) f.get().getRight()) { + LOG.info("Retry step failed for task : " + taskName); + retryFailedTasks.add(taskName); + } else { + taskFailureCounter.get(taskName).set(0); + storeLastCompletedTransaction(taskName, events.getLastEventInfo()); + } + } + } + + //Reprocess + if (!retryFailedTasks.isEmpty()) { Review comment: This is a very heavy operation, add a TODO for queuing up reprocess tasks so they can be executed separately. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: common-issues-h...@hadoop.apache.org
