FANNG1 commented on code in PR #8607: URL: https://github.com/apache/gravitino/pull/8607#discussion_r2362185223
########## core/src/main/java/org/apache/gravitino/job/local/SparkProcessBuilder.java: ########## @@ -0,0 +1,129 @@ +/* + * 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.gravitino.job.local; + +import static org.apache.gravitino.job.local.LocalJobExecutorConfigs.SPARK_HOME; + +import com.google.common.collect.Lists; +import java.io.File; +import java.util.List; +import java.util.Map; +import org.apache.arrow.util.Preconditions; +import org.apache.arrow.util.VisibleForTesting; +import org.apache.commons.lang3.StringUtils; +import org.apache.gravitino.job.SparkJobTemplate; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** + * The SparkProcessBuilder class is responsible for constructing and starting a local Spark job + * process using the spark-submit command. Using local job executor to run Spark job has some + * limitations: 1. It uses the aliveness of the spark-submit process to determine if the job is + * still running, which is not accurate for cluster deploy mode. 2. It cannot support user + * impersonation for now. + */ +public class SparkProcessBuilder extends LocalProcessBuilder { + + private static final Logger LOG = LoggerFactory.getLogger(SparkProcessBuilder.class); + + private final String sparkSubmit; + + protected SparkProcessBuilder(SparkJobTemplate sparkJobTemplate, Map<String, String> configs) { + super(sparkJobTemplate, configs); + String sparkHome = configs.get(SPARK_HOME); + Preconditions.checkArgument( + StringUtils.isNotBlank(sparkHome), + "gravitino.jobExecutor.local.sparkHome must be set for Spark jobs"); + + this.sparkSubmit = sparkHome + "/bin/spark-submit"; + File sparkSubmitFile = new File(sparkSubmit); + Preconditions.checkArgument( + sparkSubmitFile.canExecute(), + "spark-submit is not found or not executable: " + sparkSubmit); + } + + @VisibleForTesting + static List<String> generateSparkSubmitCommand( + String sparkSubmit, SparkJobTemplate sparkJobTemplate) { + List<String> commandList = Lists.newArrayList(sparkSubmit); + + // Add the main class + if (StringUtils.isNotBlank(sparkJobTemplate.className())) { + commandList.add("--class"); + commandList.add(sparkJobTemplate.className()); + } + + // Add jars if it is not empty + if (!sparkJobTemplate.jars().isEmpty()) { + commandList.add("--jars"); + commandList.add(String.join(",", sparkJobTemplate.jars())); + } + + // Add files if it is not empty + if (!sparkJobTemplate.files().isEmpty()) { + commandList.add("--files"); + commandList.add(String.join(",", sparkJobTemplate.files())); + } + + // Add archives if it is not empty + if (!sparkJobTemplate.archives().isEmpty()) { + commandList.add("--archives"); + commandList.add(String.join(",", sparkJobTemplate.archives())); + } + + // Add the Spark configs + for (Map.Entry<String, String> entry : sparkJobTemplate.configs().entrySet()) { + commandList.add("--conf"); + commandList.add(entry.getKey() + "=" + entry.getValue()); + } + + // Add the main executable + commandList.add(sparkJobTemplate.executable()); + + // Add the arguments + commandList.addAll(sparkJobTemplate.arguments()); + + return commandList; + } + + @Override + public Process start() { + SparkJobTemplate sparkJobTemplate = (SparkJobTemplate) jobTemplate; + List<String> commandList = generateSparkSubmitCommand(sparkSubmit, sparkJobTemplate); + + ProcessBuilder builder = new ProcessBuilder(commandList); + builder.directory(workingDirectory); + builder.environment().putAll(sparkJobTemplate.environments()); + + File outputFile = new File(workingDirectory, "output.log"); + File errorFile = new File(workingDirectory, "error.log"); + + builder.redirectOutput(outputFile); + builder.redirectError(errorFile); + + LOG.info("Starting local Spark job with command: {}", String.join(" ", commandList)); Review Comment: Could we add environment vars to the logs to help debug? -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
