big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

HDFS-9637. Tests for RollingFileSystemSink. (Daniel Templeton via kasha)

Browse Source
This commit is contained in:
Karthik Kambatla 2016-02-10 10:08:05 -08:00
parent 4ef1324ff6
commit bc425a623f
2 changed files with 289 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
View File

@ -986,6 +986,8 @@ Release 2.9.0 - UNRELEASED
HDFS-7764. DirectoryScanner shouldn't abort the scan if one directory had
an error (Rakesh R via cmccabe)
HDFS-9637. Tests for RollingFileSystemSink. (Daniel Templeton via kasha)
OPTIMIZATIONS
BUG FIXES

287
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/metrics2/sink/TestRollingFileSystemSinkWithHdfs.java Normal file
View File

@ -0,0 +1,287 @@
/*
* 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.metrics2.sink;
import java.io.IOException;
import java.net.URI;
import java.util.Calendar;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.metrics2.MetricsSystem;
import org.apache.hadoop.metrics2.sink.RollingFileSystemSinkTestBase.MyMetrics1;
import org.junit.After;
import org.junit.Test;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertTrue;
import org.junit.Before;
/**
* Test the {@link RollingFileSystemSink} class in the context of HDFS.
*/
public class TestRollingFileSystemSinkWithHdfs
extends RollingFileSystemSinkTestBase {
private static final int NUM_DATANODES = 4;
private MiniDFSCluster cluster;
/**
* Create a {@link MiniDFSCluster} instance with four nodes. The
* node count is required to allow append to function. Also clear the
* sink's test flags.
*
* @throws IOException thrown if cluster creation fails
*/
@Before
public void setupHdfs() throws IOException {
Configuration conf = new Configuration();
// It appears that since HDFS-265, append is always enabled.
cluster =
new MiniDFSCluster.Builder(conf).numDataNodes(NUM_DATANODES).build();
// Also clear sink flags
RollingFileSystemSink.isTest = false;
RollingFileSystemSink.hasFlushed = false;
}
/**
* Stop the {@link MiniDFSCluster}.
*/
@After
public void shutdownHdfs() {
if (cluster != null) {
cluster.shutdown();
}
}
/**
* Test writing logs to HDFS.
*
* @throws Exception thrown when things break
*/
@Test
public void testWrite() throws Exception {
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, false, true);
assertMetricsContents(doWriteTest(ms, path, 1));
}
/**
* Test writing logs to HDFS if append is enabled and the log file already
* exists.
*
* @throws Exception thrown when things break
*/
@Test
public void testAppend() throws Exception {
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
assertExtraContents(doAppendTest(path, false, true, 1));
}
/**
* Test writing logs to HDFS if append is enabled, the log file already
* exists, and the sink is set to ignore errors.
*
* @throws Exception thrown when things break
*/
@Test
public void testSilentAppend() throws Exception {
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
assertExtraContents(doAppendTest(path, false, true, 1));
}
/**
* Test writing logs to HDFS without append enabled, when the log file already
* exists.
*
* @throws Exception thrown when things break
*/
@Test
public void testNoAppend() throws Exception {
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
assertMetricsContents(doAppendTest(path, false, false, 2));
}
/**
* Test writing logs to HDFS without append enabled, with ignore errors
* enabled, and when the log file already exists.
*
* @throws Exception thrown when things break
*/
@Test
public void testSilentOverwrite() throws Exception {
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
assertMetricsContents(doAppendTest(path, true, false, 2));
}
/**
* Test that writing to HDFS fails when HDFS is unavailable.
*
* @throws IOException thrown when reading or writing log files
*/
@Test
public void testFailedWrite() throws IOException {
final String path =
"hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, false, false);
new MyMetrics1().registerWith(ms);
shutdownHdfs();
ErrorSink.errored = false;
ms.publishMetricsNow(); // publish the metrics
assertTrue("No exception was generated while writing metrics "
+ "even though HDFS was unavailable", ErrorSink.errored);
ms.stop();
ms.shutdown();
}
/**
* Test that closing a file in HDFS fails when HDFS is unavailable.
*
* @throws IOException thrown when reading or writing log files
*/
@Test
public void testFailedClose() throws IOException {
final String path =
"hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, false, false);
new MyMetrics1().registerWith(ms);
ms.publishMetricsNow(); // publish the metrics
shutdownHdfs();
ErrorSink.errored = false;
ms.stop();
assertTrue("No exception was generated while stopping sink "
+ "even though HDFS was unavailable", ErrorSink.errored);
ms.shutdown();
}
/**
* Test that writing to HDFS fails silently when HDFS is unavailable.
*
* @throws IOException thrown when reading or writing log files
* @throws java.lang.InterruptedException thrown if interrupted
*/
@Test
public void testSilentFailedWrite() throws IOException, InterruptedException {
final String path =
"hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, true, false);
new MyMetrics1().registerWith(ms);
shutdownHdfs();
ErrorSink.errored = false;
ms.publishMetricsNow(); // publish the metrics
assertFalse("An exception was generated writing metrics "
+ "while HDFS was unavailable, even though the sink is set to "
+ "ignore errors", ErrorSink.errored);
ms.stop();
ms.shutdown();
}
/**
* Test that closing a file in HDFS silently fails when HDFS is unavailable.
*
* @throws IOException thrown when reading or writing log files
*/
@Test
public void testSilentFailedClose() throws IOException {
final String path =
"hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, true, false);
new MyMetrics1().registerWith(ms);
ms.publishMetricsNow(); // publish the metrics
shutdownHdfs();
ErrorSink.errored = false;
ms.stop();
assertFalse("An exception was generated stopping sink "
+ "while HDFS was unavailable, even though the sink is set to "
+ "ignore errors", ErrorSink.errored);
ms.shutdown();
}
/**
* This test specifically checks whether the flusher thread is automatically
* flushing the files. It unfortunately can only test with the alternative
* flushing schedule (because of test timing), but it's better than nothing.
*
* @throws Exception thrown if something breaks
*/
@Test
public void testFlushThread() throws Exception {
RollingFileSystemSink.isTest = true;
RollingFileSystemSink.hasFlushed = false;
String path = "hdfs://" + cluster.getNameNode().getHostAndPort() + "/tmp";
MetricsSystem ms = initMetricsSystem(path, true, false);
new MyMetrics1().registerWith(ms);
// Publish the metrics
ms.publishMetricsNow();
// Pubish again because the first write seems to get properly flushed
// regardless.
ms.publishMetricsNow();
// Sleep until the flusher has run
while (!RollingFileSystemSink.hasFlushed) {
Thread.sleep(50L);
}
Calendar now = getNowNotTopOfHour();
FileSystem fs = FileSystem.newInstance(new URI(path), new Configuration());
Path currentDir = new Path(path, DATE_FORMAT.format(now.getTime()));
Path currentFile =
findMostRecentLogFile(fs, new Path(currentDir, getLogFilename()));
FileStatus status = fs.getFileStatus(currentFile);
// Each metrics record is 118+ bytes, depending on hostname
assertTrue("The flusher thread didn't flush the log contents. Expected "
+ "at least 236 bytes in the log file, but got " + status.getLen(),
status.getLen() >= 236);
ms.stop();
}
}