From 0a33a98eb4bbfb07392a549c42d0830a14d52d66 Mon Sep 17 00:00:00 2001 From: Arpit Agarwal Date: Fri, 13 Feb 2015 11:56:25 -0800 Subject: [PATCH] HDFS-7776. Adding additional unit tests for Quota By Storage Type. (Contributed by Xiaoyu Yao) --- hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt | 21 +- .../namenode/TestQuotaByStorageType.java | 187 +++++++++++++++++- 2 files changed, 199 insertions(+), 9 deletions(-) diff --git a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt index 6463010392c..7705f878050 100644 --- a/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt +++ b/hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt @@ -292,9 +292,6 @@ Trunk (Unreleased) HDFS-7721. The HDFS BlockScanner may run fast during the first hour (cmccabe) - HDFS-7751. Fix TestHDFSCLI for quota with storage type. (Xiaoyu Yao - via szetszwo) - HDFS-7670. HDFS Quota guide has typos, incomplete command lines (Brahma Reddy Battula via aw) @@ -340,8 +337,8 @@ Release 2.7.0 - UNRELEASED HDFS-6133. Add a feature for replica pinning so that a pinned replica will not be moved by Balancer/Mover. (zhaoyunjiong via szetszwo) - HDFS-7584. Enable Quota Support for Storage Types. (Xiaoyu Yao via - Arpit Agarwal) + HDFS-7584. Enable Quota Support for Storage Types (See breakdown of + tasks below) IMPROVEMENTS @@ -948,6 +945,20 @@ Release 2.7.0 - UNRELEASED HDFS-7704. DN heartbeat to Active NN may be blocked and expire if connection to Standby NN continues to time out (Rushabh Shah via kihwal) + BREAKDOWN OF HDFS-7584 SUBTASKS AND RELATED JIRAS + + HDFS-7720. Quota by Storage Type API, tools and ClientNameNode + Protocol changes. (Xiaoyu Yao via Arpit Agarwal) + + HDFS-7723. Quota By Storage Type namenode implemenation. (Xiaoyu Yao + via Arpit Agarwal) + + HDFS-7751. Fix TestHDFSCLI for quota with storage type. (Xiaoyu Yao + via szetszwo) + + HDFS-7776. Adding additional unit tests for Quota By Storage Type. + (Xiaoyu Yao via Arpit Agarwal) + Release 2.6.1 - UNRELEASED INCOMPATIBLE CHANGES diff --git a/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestQuotaByStorageType.java b/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestQuotaByStorageType.java index b0d5d87b8ce..57f026d9635 100644 --- a/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestQuotaByStorageType.java +++ b/hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestQuotaByStorageType.java @@ -19,7 +19,6 @@ package org.apache.hadoop.hdfs.server.namenode; import static org.apache.hadoop.hdfs.StorageType.DEFAULT; import static org.junit.Assert.assertEquals; - import static org.junit.Assert.assertFalse; import static org.junit.Assert.assertTrue; import static org.junit.Assert.fail; @@ -30,15 +29,18 @@ package org.apache.hadoop.hdfs.server.namenode; import org.apache.hadoop.hdfs.*; import org.apache.hadoop.hdfs.protocol.HdfsConstants; import org.apache.hadoop.hdfs.server.namenode.snapshot.SnapshotTestHelper; + import org.apache.hadoop.test.GenericTestUtils; import org.junit.After; import org.junit.Before; import org.junit.Test; + import java.io.IOException; + public class TestQuotaByStorageType { private static final int BLOCKSIZE = 1024; private static final short REPLICATION = 3; - static final long seed = 0L; + private static final long seed = 0L; private static final Path dir = new Path("/TestQuotaByStorageType"); private Configuration conf; @@ -219,7 +221,6 @@ public class TestQuotaByStorageType { // Verify space consumed and remaining quota long ssdConsumed = fnode.asDirectory().getDirectoryWithQuotaFeature() .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); - ; assertEquals(file1Len, ssdConsumed); // move file from foo to bar @@ -356,7 +357,6 @@ public class TestQuotaByStorageType { assertTrue(fnode.isQuotaSet()); long currentSSDConsumed = fnode.asDirectory().getDirectoryWithQuotaFeature() .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); - ; assertEquals(file1Len, currentSSDConsumed); // Create the 2nd file of size BLOCKSIZE under child directory and expect quota exceeded exception @@ -399,11 +399,19 @@ public class TestQuotaByStorageType { } } + /** + * Both traditional space quota and the storage type quota for SSD are set and + * not exceeded. + */ @Test(timeout = 60000) public void testQuotaByStorageTypeWithTraditionalQuota() throws Exception { final Path foo = new Path(dir, "foo"); dfs.mkdirs(foo); + + dfs.setStoragePolicy(foo, HdfsConstants.ONESSD_STORAGE_POLICY_NAME); + dfs.setQuotaByStorageType(foo, StorageType.SSD, BLOCKSIZE * 10); dfs.setQuota(foo, Long.MAX_VALUE - 1, REPLICATION * BLOCKSIZE * 10); + INode fnode = fsdir.getINode4Write(foo.toString()); assertTrue(fnode.isDirectory()); assertTrue(fnode.isQuotaSet()); @@ -434,6 +442,76 @@ public class TestQuotaByStorageType { counts.getDiskSpace()); } + /** + * Both traditional space quota and the storage type quota for SSD are set and + * exceeded. expect DSQuotaExceededException is thrown as we check traditional + * space quota first and then storage type quota. + */ + @Test(timeout = 60000) + public void testQuotaByStorageTypeAndTraditionalQuotaException1() + throws Exception { + testQuotaByStorageTypeOrTraditionalQuotaExceededCase( + 4 * REPLICATION, 4, 5, REPLICATION); + } + + /** + * Both traditional space quota and the storage type quota for SSD are set and + * SSD quota is exceeded but traditional space quota is not exceeded. + */ + @Test(timeout = 60000) + public void testQuotaByStorageTypeAndTraditionalQuotaException2() + throws Exception { + testQuotaByStorageTypeOrTraditionalQuotaExceededCase( + 5 * REPLICATION, 4, 5, REPLICATION); + } + + /** + * Both traditional space quota and the storage type quota for SSD are set and + * traditional space quota is exceeded but SSD quota is not exceeded. + */ + @Test(timeout = 60000) + public void testQuotaByStorageTypeAndTraditionalQuotaException3() + throws Exception { + testQuotaByStorageTypeOrTraditionalQuotaExceededCase( + 4 * REPLICATION, 5, 5, REPLICATION); + } + + private void testQuotaByStorageTypeOrTraditionalQuotaExceededCase( + long storageSpaceQuotaInBlocks, long ssdQuotaInBlocks, + long testFileLenInBlocks, short replication) throws Exception { + final String METHOD_NAME = GenericTestUtils.getMethodName(); + final Path testDir = new Path(dir, METHOD_NAME); + + dfs.mkdirs(testDir); + dfs.setStoragePolicy(testDir, HdfsConstants.ONESSD_STORAGE_POLICY_NAME); + + final long ssdQuota = BLOCKSIZE * ssdQuotaInBlocks; + final long storageSpaceQuota = BLOCKSIZE * storageSpaceQuotaInBlocks; + + dfs.setQuota(testDir, Long.MAX_VALUE - 1, storageSpaceQuota); + dfs.setQuotaByStorageType(testDir, StorageType.SSD, ssdQuota); + + INode testDirNode = fsdir.getINode4Write(testDir.toString()); + assertTrue(testDirNode.isDirectory()); + assertTrue(testDirNode.isQuotaSet()); + + Path createdFile = new Path(testDir, "created_file.data"); + long fileLen = testFileLenInBlocks * BLOCKSIZE; + + try { + DFSTestUtil.createFile(dfs, createdFile, BLOCKSIZE / 16, + fileLen, BLOCKSIZE, replication, seed); + fail("Should have failed with DSQuotaExceededException or " + + "QuotaByStorageTypeExceededException "); + } catch (Throwable t) { + LOG.info("Got expected exception ", t); + long currentSSDConsumed = testDirNode.asDirectory().getDirectoryWithQuotaFeature() + .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); + assertEquals(Math.min(ssdQuota, storageSpaceQuota/replication), + currentSSDConsumed); + } + } + @Test(timeout = 60000) public void testQuotaByStorageTypeWithSnapshot() throws Exception { final Path sub1 = new Path(dir, "Sub1"); @@ -521,4 +599,105 @@ public class TestQuotaByStorageType { .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); assertEquals(newFile1Len, ssdConsumed); } + + @Test + public void testQuotaByStorageTypePersistenceInEditLog() throws IOException { + final String METHOD_NAME = GenericTestUtils.getMethodName(); + final Path testDir = new Path(dir, METHOD_NAME); + Path createdFile1 = new Path(testDir, "created_file1.data"); + dfs.mkdirs(testDir); + + // set storage policy on testDir to ONESSD + dfs.setStoragePolicy(testDir, HdfsConstants.ONESSD_STORAGE_POLICY_NAME); + + // set quota by storage type on testDir + final long SSD_QUOTA = BLOCKSIZE * 4; + dfs.setQuotaByStorageType(testDir, StorageType.SSD, SSD_QUOTA); + INode testDirNode = fsdir.getINode4Write(testDir.toString()); + assertTrue(testDirNode.isDirectory()); + assertTrue(testDirNode.isQuotaSet()); + + // Create file of size 2 * BLOCKSIZE under testDir + long file1Len = BLOCKSIZE * 2; + int bufLen = BLOCKSIZE / 16; + DFSTestUtil.createFile(dfs, createdFile1, bufLen, file1Len, BLOCKSIZE, REPLICATION, seed); + + // Verify SSD consumed before namenode restart + long ssdConsumed = testDirNode.asDirectory().getDirectoryWithQuotaFeature() + .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); + assertEquals(file1Len, ssdConsumed); + + // Restart namenode to make sure the editlog is correct + cluster.restartNameNode(true); + + INode testDirNodeAfterNNRestart = fsdir.getINode4Write(testDir.toString()); + // Verify quota is still set + assertTrue(testDirNode.isDirectory()); + assertTrue(testDirNode.isQuotaSet()); + + QuotaCounts qc = testDirNodeAfterNNRestart.getQuotaCounts(); + assertEquals(SSD_QUOTA, qc.getTypeSpace(StorageType.SSD)); + for (StorageType t: StorageType.getTypesSupportingQuota()) { + if (t != StorageType.SSD) { + assertEquals(HdfsConstants.QUOTA_RESET, qc.getTypeSpace(t)); + } + } + + long ssdConsumedAfterNNRestart = testDirNodeAfterNNRestart.asDirectory() + .getDirectoryWithQuotaFeature() + .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); + assertEquals(file1Len, ssdConsumedAfterNNRestart); + } + + @Test + public void testQuotaByStorageTypePersistenceInFsImage() throws IOException { + final String METHOD_NAME = GenericTestUtils.getMethodName(); + final Path testDir = new Path(dir, METHOD_NAME); + Path createdFile1 = new Path(testDir, "created_file1.data"); + dfs.mkdirs(testDir); + + // set storage policy on testDir to ONESSD + dfs.setStoragePolicy(testDir, HdfsConstants.ONESSD_STORAGE_POLICY_NAME); + + // set quota by storage type on testDir + final long SSD_QUOTA = BLOCKSIZE * 4; + dfs.setQuotaByStorageType(testDir, StorageType.SSD, SSD_QUOTA); + INode testDirNode = fsdir.getINode4Write(testDir.toString()); + assertTrue(testDirNode.isDirectory()); + assertTrue(testDirNode.isQuotaSet()); + + // Create file of size 2 * BLOCKSIZE under testDir + long file1Len = BLOCKSIZE * 2; + int bufLen = BLOCKSIZE / 16; + DFSTestUtil.createFile(dfs, createdFile1, bufLen, file1Len, BLOCKSIZE, REPLICATION, seed); + + // Verify SSD consumed before namenode restart + long ssdConsumed = testDirNode.asDirectory().getDirectoryWithQuotaFeature() + .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); + assertEquals(file1Len, ssdConsumed); + + // Restart the namenode with checkpoint to make sure fsImage is correct + dfs.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_ENTER); + dfs.saveNamespace(); + dfs.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); + cluster.restartNameNode(true); + + INode testDirNodeAfterNNRestart = fsdir.getINode4Write(testDir.toString()); + assertTrue(testDirNode.isDirectory()); + assertTrue(testDirNode.isQuotaSet()); + + QuotaCounts qc = testDirNodeAfterNNRestart.getQuotaCounts(); + assertEquals(SSD_QUOTA, qc.getTypeSpace(StorageType.SSD)); + for (StorageType t: StorageType.getTypesSupportingQuota()) { + if (t != StorageType.SSD) { + assertEquals(HdfsConstants.QUOTA_RESET, qc.getTypeSpace(t)); + } + } + + long ssdConsumedAfterNNRestart = testDirNodeAfterNNRestart.asDirectory() + .getDirectoryWithQuotaFeature() + .getSpaceConsumed().getTypeSpaces().get(StorageType.SSD); + assertEquals(file1Len, ssdConsumedAfterNNRestart); + + } } \ No newline at end of file