org.apache.hadoop.hive.service.HiveServer Maven / Gradle / Ivy
/**
* 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.hadoop.hive.service;
import java.io.BufferedReader;
import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.FileReader;
import java.io.IOException;
import java.io.PrintStream;
import java.io.UnsupportedEncodingException;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.Properties;
import org.apache.commons.cli.OptionBuilder;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.hive.common.ServerUtils;
import org.apache.hadoop.hive.common.LogUtils;
import org.apache.hadoop.hive.common.LogUtils.LogInitializationException;
import org.apache.hadoop.hive.common.cli.CommonCliOptions;
import org.apache.hadoop.hive.conf.HiveConf;
import org.apache.hadoop.hive.metastore.HiveMetaStore;
import org.apache.hadoop.hive.metastore.api.MetaException;
import org.apache.hadoop.hive.metastore.api.Schema;
import org.apache.hadoop.hive.metastore.TServerSocketKeepAlive;
import org.apache.hadoop.hive.ql.CommandNeedRetryException;
import org.apache.hadoop.hive.ql.Driver;
import org.apache.hadoop.hive.ql.plan.api.QueryPlan;
import org.apache.hadoop.hive.ql.processors.CommandProcessor;
import org.apache.hadoop.hive.ql.processors.CommandProcessorFactory;
import org.apache.hadoop.hive.ql.processors.CommandProcessorResponse;
import org.apache.hadoop.hive.ql.session.SessionState;
import org.apache.hadoop.hive.shims.ShimLoader;
import org.apache.hadoop.mapred.ClusterStatus;
import org.apache.thrift.TException;
import org.apache.thrift.TProcessor;
import org.apache.thrift.TProcessorFactory;
import org.apache.thrift.protocol.TBinaryProtocol;
import org.apache.thrift.server.TServer;
import org.apache.thrift.server.TThreadPoolServer;
import org.apache.thrift.transport.TServerSocket;
import org.apache.thrift.transport.TServerTransport;
import org.apache.thrift.transport.TTransport;
import org.apache.thrift.transport.TTransportFactory;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import com.facebook.fb303.fb_status;
/**
* Thrift Hive Server Implementation.
*/
public class HiveServer extends ThriftHive {
private static final String VERSION = "1";
/**
* default port on which to start the Hive server
*/
private static final int DEFAULT_HIVE_SERVER_PORT = 10000;
/**
* default minimum number of threads serving the Hive server
*/
private static final int DEFAULT_MIN_WORKER_THREADS = 100;
/**
* default maximum number of threads serving the Hive server
*/
private static final int DEFAULT_MAX_WORKER_THREADS = Integer.MAX_VALUE;
/**
* Handler which implements the Hive Interface This class can be used in lieu
* of the HiveClient class to get an embedded server.
*/
public static class HiveServerHandler extends HiveMetaStore.HMSHandler
implements HiveInterface {
/**
* Hive server uses org.apache.hadoop.hive.ql.Driver for run() and
* getResults() methods.
* It is the instance of the last Hive query.
*/
private Driver driver;
private CommandProcessorResponse response;
/**
* For processors other than Hive queries (Driver), they output to session.out (a temp file)
* first and the fetchOne/fetchN/fetchAll functions get the output from pipeIn.
*/
private BufferedReader pipeIn;
/**
* Flag that indicates whether the last executed command was a Hive query.
*/
private boolean isHiveQuery;
public static final Log LOG = LogFactory.getLog(HiveServer.class.getName());
/**
* Construct a new handler.
*
* @throws MetaException unable to create metastore
*/
public HiveServerHandler() throws MetaException {
this(new HiveConf(SessionState.class));
}
/**
* Construct a new handler with the specified hive configuration.
*
* @param conf caller specified hive configuration
* @throws MetaException unable to create metastore
*/
public HiveServerHandler(HiveConf conf) throws MetaException {
super(HiveServer.class.getName(), conf);
isHiveQuery = false;
driver = null;
SessionState session = new SessionState(conf);
SessionState.start(session);
setupSessionIO(session);
}
private void setupSessionIO(SessionState session) {
try {
LOG.info("Putting temp output to file " + session.getTmpOutputFile().toString());
session.in = null; // hive server's session input stream is not used
// open a per-session file in auto-flush mode for writing temp results
session.out = new PrintStream(new FileOutputStream(session.getTmpOutputFile()), true, "UTF-8");
// TODO: for hadoop jobs, progress is printed out to session.err,
// we should find a way to feed back job progress to client
session.err = new PrintStream(System.err, true, "UTF-8");
} catch (IOException e) {
LOG.error("Error in creating temp output file ", e);
try {
session.in = null;
session.out = new PrintStream(System.out, true, "UTF-8");
session.err = new PrintStream(System.err, true, "UTF-8");
} catch (UnsupportedEncodingException ee) {
ee.printStackTrace();
session.out = null;
session.err = null;
}
}
}
/**
* Executes a query.
*
* @param cmd
* HiveQL query to execute
*/
public void execute(String cmd) throws HiveServerException, TException {
HiveServerHandler.LOG.info("Running the query: " + cmd);
SessionState session = SessionState.get();
String cmd_trimmed = cmd.trim();
String[] tokens = cmd_trimmed.split("\\s");
String cmd_1 = cmd_trimmed.substring(tokens[0].length()).trim();
int ret = 0;
String errorMessage = "";
String SQLState = null;
try {
// Close the existing driver object (CommandProcessor) before creating
// the new driver (CommandProcessor) object to clean-up the resources
if (driver != null) {
driver.close();
driver = null;
}
CommandProcessor proc = CommandProcessorFactory.get(tokens[0]);
if (proc != null) {
if (proc instanceof Driver) {
isHiveQuery = true;
driver = (Driver) proc;
// In Hive server mode, we are not able to retry in the FetchTask
// case, when calling fetch quueries since execute() has returned.
// For now, we disable the test attempts.
driver.setTryCount(Integer.MAX_VALUE);
response = driver.run(cmd);
} else {
isHiveQuery = false;
driver = null;
// need to reset output for each non-Hive query
setupSessionIO(session);
response = proc.run(cmd_1);
}
ret = response.getResponseCode();
SQLState = response.getSQLState();
errorMessage = response.getErrorMessage();
}
} catch (Exception e) {
HiveServerException ex = new HiveServerException();
ex.setMessage("Error running query: " + e.toString());
ex.setErrorCode(ret == 0? -10000: ret);
throw ex;
}
if (ret != 0) {
throw new HiveServerException("Query returned non-zero code: " + ret
+ ", cause: " + errorMessage, ret, SQLState);
}
}
/**
* Should be called by the client at the end of a session.
*/
public void clean() {
if (driver != null) {
driver.close();
driver.destroy();
}
SessionState session = SessionState.get();
if (session.getTmpOutputFile() != null) {
session.getTmpOutputFile().delete();
}
pipeIn = null;
}
/**
* Return the status information about the Map-Reduce cluster.
*/
public HiveClusterStatus getClusterStatus() throws HiveServerException,
TException {
HiveClusterStatus hcs;
try {
Driver drv = new Driver();
drv.init();
ClusterStatus cs = drv.getClusterStatus();
JobTrackerState state = JobTrackerState.valueOf(ShimLoader.getHadoopShims().getJobTrackerState(cs).name());
hcs = new HiveClusterStatus(cs.getTaskTrackers(), cs.getMapTasks(), cs
.getReduceTasks(), cs.getMaxMapTasks(), cs.getMaxReduceTasks(),
state);
} catch (Exception e) {
LOG.error(e.toString());
e.printStackTrace();
HiveServerException ex = new HiveServerException();
ex.setMessage("Unable to get cluster status: " + e.toString());
throw ex;
}
return hcs;
}
/**
* Return the Hive schema of the query result.
*/
public Schema getSchema() throws HiveServerException, TException {
if (!isHiveQuery) {
Schema schema = response.getSchema();
if (schema == null) {
// Return empty schema if the last command was not a Hive query
return new Schema();
}
else {
return schema;
}
}
assert driver != null: "getSchema() is called on a Hive query and driver is NULL.";
try {
Schema schema = driver.getSchema();
if (schema == null) {
schema = new Schema();
}
LOG.info("Returning schema: " + schema);
return schema;
} catch (Exception e) {
LOG.error(e.toString());
e.printStackTrace();
HiveServerException ex = new HiveServerException();
ex.setMessage("Unable to get schema: " + e.toString());
throw ex;
}
}
/**
* Return the Thrift schema of the query result.
*/
public Schema getThriftSchema() throws HiveServerException, TException {
if (!isHiveQuery) {
// Return empty schema if the last command was not a Hive query
return new Schema();
}
assert driver != null: "getThriftSchema() is called on a Hive query and driver is NULL.";
try {
Schema schema = driver.getThriftSchema();
if (schema == null) {
schema = new Schema();
}
LOG.info("Returning schema: " + schema);
return schema;
} catch (Exception e) {
LOG.error(e.toString());
e.printStackTrace();
HiveServerException ex = new HiveServerException();
ex.setMessage("Unable to get schema: " + e.toString());
throw ex;
}
}
/**
* Fetches the next row in a query result set.
*
* @return the next row in a query result set. null if there is no more row
* to fetch.
*/
public String fetchOne() throws HiveServerException, TException {
if (!isHiveQuery) {
// Return no results if the last command was not a Hive query
List results = new ArrayList(1);
readResults(results, 1);
if (results.size() > 0) {
return results.get(0);
} else { // throw an EOF exception
throw new HiveServerException("OK", 0, "");
}
}
assert driver != null: "fetchOne() is called on a Hive query and driver is NULL.";
ArrayList result = new ArrayList();
driver.setMaxRows(1);
try {
if (driver.getResults(result)) {
return result.get(0);
}
// TODO: Cannot return null here because thrift cannot handle nulls
// TODO: Returning empty string for now. Need to figure out how to
// TODO: return null in some other way
throw new HiveServerException("OK", 0, "");
// return "";
} catch (CommandNeedRetryException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
} catch (IOException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
}
}
private void cleanTmpFile() {
if (pipeIn != null) {
SessionState session = SessionState.get();
File tmp = session.getTmpOutputFile();
tmp.delete();
pipeIn = null;
}
}
/**
* Reads the temporary results for non-Hive (non-Driver) commands to the
* resulting List of strings.
* @param results list of strings containing the results
* @param nLines number of lines read at once. If it is <= 0, then read all lines.
*/
private void readResults(List results, int nLines) {
if (pipeIn == null) {
SessionState session = SessionState.get();
File tmp = session.getTmpOutputFile();
try {
pipeIn = new BufferedReader(new FileReader(tmp));
} catch (FileNotFoundException e) {
LOG.error("File " + tmp + " not found. ", e);
return;
}
}
boolean readAll = false;
for (int i = 0; i < nLines || nLines <= 0; ++i) {
try {
String line = pipeIn.readLine();
if (line == null) {
// reached the end of the result file
readAll = true;
break;
} else {
results.add(line);
}
} catch (IOException e) {
LOG.error("Reading temp results encountered an exception: ", e);
readAll = true;
}
}
if (readAll) {
cleanTmpFile();
}
}
/**
* Fetches numRows rows.
*
* @param numRows
* Number of rows to fetch.
* @return A list of rows. The size of the list is numRows if there are at
* least numRows rows available to return. The size is smaller than
* numRows if there aren't enough rows. The list will be empty if
* there is no more row to fetch or numRows == 0.
* @throws HiveServerException
* Invalid value for numRows (numRows < 0)
*/
public List fetchN(int numRows) throws HiveServerException,
TException {
if (numRows < 0) {
HiveServerException ex = new HiveServerException();
ex.setMessage("Invalid argument for number of rows: " + numRows);
throw ex;
}
ArrayList result = new ArrayList();
if (!isHiveQuery) {
readResults(result, numRows);
return result;
}
assert driver != null: "fetchN() is called on a Hive query and driver is NULL.";
driver.setMaxRows(numRows);
try {
driver.getResults(result);
} catch (CommandNeedRetryException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
} catch (IOException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
}
return result;
}
/**
* Fetches all the rows in a result set.
*
* @return All the rows in a result set of a query executed using execute
* method.
*
* TODO: Currently the server buffers all the rows before returning
* them to the client. Decide whether the buffering should be done
* in the client.
*/
public List fetchAll() throws HiveServerException, TException {
ArrayList rows = new ArrayList();
ArrayList result = new ArrayList();
if (!isHiveQuery) {
// Return all results if numRows <= 0
readResults(result, 0);
return result;
}
try {
while (driver.getResults(result)) {
rows.addAll(result);
result.clear();
}
} catch (CommandNeedRetryException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
} catch (IOException e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.getMessage());
throw ex;
}
return rows;
}
/**
* Return the status of the server.
*/
@Override
public fb_status getStatus() {
return fb_status.ALIVE;
}
/**
* Return the version of the server software.
*/
@Override
public String getVersion() {
return VERSION;
}
@Override
public QueryPlan getQueryPlan() throws HiveServerException, TException {
QueryPlan qp = new QueryPlan();
if (!isHiveQuery) {
return qp;
}
assert driver != null: "getQueryPlan() is called on a Hive query and driver is NULL.";
// TODO for now only return one query at a time
// going forward, all queries associated with a single statement
// will be returned in a single QueryPlan
try {
qp.addToQueries(driver.getQueryPlan());
} catch (Exception e) {
HiveServerException ex = new HiveServerException();
ex.setMessage(e.toString());
throw ex;
}
return qp;
}
}
/**
* ThriftHiveProcessorFactory.
*
*/
public static class ThriftHiveProcessorFactory extends TProcessorFactory {
private final HiveConf conf;
public ThriftHiveProcessorFactory(TProcessor processor, HiveConf conf) {
super(processor);
this.conf = conf;
}
@Override
@SuppressWarnings("unchecked")
public TProcessor getProcessor(TTransport trans) {
try {
Iface handler = new HiveServerHandler(new HiveConf(conf));
return new ThriftHive.Processor(handler);
} catch (Exception e) {
HiveServerHandler.LOG.warn("Failed to get processor by exception " + e, e);
trans.close();
throw new RuntimeException(e);
}
}
}
/**
* HiveServer specific CLI
*
*/
static public class HiveServerCli extends CommonCliOptions {
private static final String OPTION_MAX_WORKER_THREADS = "maxWorkerThreads";
private static final String OPTION_MIN_WORKER_THREADS = "minWorkerThreads";
public int port = DEFAULT_HIVE_SERVER_PORT;
public int minWorkerThreads = DEFAULT_MIN_WORKER_THREADS;
public int maxWorkerThreads = DEFAULT_MAX_WORKER_THREADS;
@SuppressWarnings("static-access")
public HiveServerCli() {
super("hiveserver", true);
// -p port
OPTIONS.addOption(OptionBuilder
.hasArg()
.withArgName("port")
.withDescription("Hive Server port number, default:"
+ DEFAULT_HIVE_SERVER_PORT)
.create('p'));
// min worker thread count
OPTIONS.addOption(OptionBuilder
.hasArg()
.withLongOpt(OPTION_MIN_WORKER_THREADS)
.withDescription("minimum number of worker threads, default:"
+ DEFAULT_MIN_WORKER_THREADS)
.create());
// max worker thread count
OPTIONS.addOption(OptionBuilder
.hasArg()
.withLongOpt(OPTION_MAX_WORKER_THREADS)
.withDescription("maximum number of worker threads, default:"
+ DEFAULT_MAX_WORKER_THREADS)
.create());
}
@Override
public void parse(String[] args) {
super.parse(args);
// support the old syntax "hiveserver [port [threads]]" but complain
args = commandLine.getArgs();
if (args.length >= 1) {
// complain about the deprecated syntax -- but still run
System.err.println(
"This usage has been deprecated, consider using the new command "
+ "line syntax (run with -h to see usage information)");
port = Integer.parseInt(args[0]);
}
if (args.length >= 2) {
minWorkerThreads = Integer.parseInt(args[1]);
}
// notice that command line options take precedence over the
// deprecated (old style) naked args...
if (commandLine.hasOption('p')) {
port = Integer.parseInt(commandLine.getOptionValue('p'));
} else {
// legacy handling
String hivePort = System.getenv("HIVE_PORT");
if (hivePort != null) {
port = Integer.parseInt(hivePort);
}
}
if (commandLine.hasOption(OPTION_MIN_WORKER_THREADS)) {
minWorkerThreads = Integer.parseInt(
commandLine.getOptionValue(OPTION_MIN_WORKER_THREADS));
}
if (commandLine.hasOption(OPTION_MAX_WORKER_THREADS)) {
maxWorkerThreads = Integer.parseInt(
commandLine.getOptionValue(OPTION_MAX_WORKER_THREADS));
}
}
}
public static void main(String[] args) {
try {
HiveServerCli cli = new HiveServerCli();
cli.parse(args);
// NOTE: It is critical to do this prior to initializing log4j, otherwise
// any log specific settings via hiveconf will be ignored
Properties hiveconf = cli.addHiveconfToSystemProperties();
// NOTE: It is critical to do this here so that log4j is reinitialized
// before any of the other core hive classes are loaded
try {
LogUtils.initHiveLog4j();
} catch (LogInitializationException e) {
HiveServerHandler.LOG.warn(e.getMessage());
}
HiveConf conf = new HiveConf(HiveServerHandler.class);
ServerUtils.cleanUpScratchDir(conf);
boolean tcpKeepAlive = conf.getBoolVar(HiveConf.ConfVars.SERVER_TCP_KEEP_ALIVE);
TServerTransport serverTransport = tcpKeepAlive ? new TServerSocketKeepAlive(cli.port) : new TServerSocket(cli.port, 1000 * conf.getIntVar(HiveConf.ConfVars.SERVER_READ_SOCKET_TIMEOUT));
// set all properties specified on the command line
for (Map.Entry item : hiveconf.entrySet()) {
conf.set((String) item.getKey(), (String) item.getValue());
}
ThriftHiveProcessorFactory hfactory =
new ThriftHiveProcessorFactory(null, conf);
TThreadPoolServer.Args sargs = new TThreadPoolServer.Args(serverTransport)
.processorFactory(hfactory)
.transportFactory(new TTransportFactory())
.protocolFactory(new TBinaryProtocol.Factory())
.minWorkerThreads(cli.minWorkerThreads)
.maxWorkerThreads(cli.maxWorkerThreads);
TServer server = new TThreadPoolServer(sargs);
String msg = "Starting hive server on port " + cli.port
+ " with " + cli.minWorkerThreads + " min worker threads and "
+ cli.maxWorkerThreads + " max worker threads";
HiveServerHandler.LOG.info(msg);
HiveServerHandler.LOG.info("TCP keepalive = " + tcpKeepAlive);
if (cli.isVerbose()) {
System.err.println(msg);
}
server.serve();
} catch (Exception x) {
x.printStackTrace();
}
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy