Search in sources :

Example 36 with CopyListingFileStatus

use of org.apache.hadoop.tools.CopyListingFileStatus in project hadoop by apache.

the class TestCopyMapper method testFailCopyWithAccessControlException.

@Test(timeout = 40000)
public void testFailCopyWithAccessControlException() {
    try {
        deleteState();
        createSourceData();
        UserGroupInformation tmpUser = UserGroupInformation.createRemoteUser("guest");
        final CopyMapper copyMapper = new CopyMapper();
        final StubContext stubContext = tmpUser.doAs(new PrivilegedAction<StubContext>() {

            @Override
            public StubContext run() {
                try {
                    return new StubContext(getConfiguration(), null, 0);
                } catch (Exception e) {
                    LOG.error("Exception encountered ", e);
                    throw new RuntimeException(e);
                }
            }
        });
        EnumSet<DistCpOptions.FileAttribute> preserveStatus = EnumSet.allOf(DistCpOptions.FileAttribute.class);
        preserveStatus.remove(DistCpOptions.FileAttribute.ACL);
        preserveStatus.remove(DistCpOptions.FileAttribute.XATTR);
        final Mapper<Text, CopyListingFileStatus, Text, Text>.Context<Text, CopyListingFileStatus, Text, Text> context = stubContext.getContext();
        context.getConfiguration().set(DistCpConstants.CONF_LABEL_PRESERVE_STATUS, DistCpUtils.packAttributes(preserveStatus));
        touchFile(SOURCE_PATH + "/src/file");
        OutputStream out = cluster.getFileSystem().create(new Path(TARGET_PATH + "/src/file"));
        out.write("hello world".getBytes());
        out.close();
        cluster.getFileSystem().setPermission(new Path(SOURCE_PATH + "/src/file"), new FsPermission(FsAction.READ, FsAction.READ, FsAction.READ));
        cluster.getFileSystem().setPermission(new Path(TARGET_PATH + "/src/file"), new FsPermission(FsAction.READ, FsAction.READ, FsAction.READ));
        final FileSystem tmpFS = tmpUser.doAs(new PrivilegedAction<FileSystem>() {

            @Override
            public FileSystem run() {
                try {
                    return FileSystem.get(configuration);
                } catch (IOException e) {
                    LOG.error("Exception encountered ", e);
                    Assert.fail("Test failed: " + e.getMessage());
                    throw new RuntimeException("Test ought to fail here");
                }
            }
        });
        tmpUser.doAs(new PrivilegedAction<Integer>() {

            @Override
            public Integer run() {
                try {
                    copyMapper.setup(context);
                    copyMapper.map(new Text("/src/file"), new CopyListingFileStatus(tmpFS.getFileStatus(new Path(SOURCE_PATH + "/src/file"))), context);
                    Assert.fail("Didn't expect the file to be copied");
                } catch (AccessControlException ignore) {
                } catch (Exception e) {
                    // itself.
                    if (e.getCause() == null || e.getCause().getCause() == null || !(e.getCause().getCause() instanceof AccessControlException)) {
                        throw new RuntimeException(e);
                    }
                }
                return null;
            }
        });
    } catch (Exception e) {
        LOG.error("Exception encountered ", e);
        Assert.fail("Test failed: " + e.getMessage());
    }
}
Also used : Path(org.apache.hadoop.fs.Path) FSDataOutputStream(org.apache.hadoop.fs.FSDataOutputStream) DataOutputStream(java.io.DataOutputStream) OutputStream(java.io.OutputStream) AccessControlException(org.apache.hadoop.security.AccessControlException) Text(org.apache.hadoop.io.Text) IOException(java.io.IOException) IOException(java.io.IOException) AccessControlException(org.apache.hadoop.security.AccessControlException) Mapper(org.apache.hadoop.mapreduce.Mapper) DistCpOptions(org.apache.hadoop.tools.DistCpOptions) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) StubContext(org.apache.hadoop.tools.StubContext) FileSystem(org.apache.hadoop.fs.FileSystem) FsPermission(org.apache.hadoop.fs.permission.FsPermission) UserGroupInformation(org.apache.hadoop.security.UserGroupInformation) Test(org.junit.Test)

Example 37 with CopyListingFileStatus

use of org.apache.hadoop.tools.CopyListingFileStatus in project hadoop by apache.

the class TestCopyMapper method doTestIgnoreFailures.

private void doTestIgnoreFailures(boolean ignoreFailures) {
    try {
        deleteState();
        createSourceData();
        FileSystem fs = cluster.getFileSystem();
        CopyMapper copyMapper = new CopyMapper();
        StubContext stubContext = new StubContext(getConfiguration(), null, 0);
        Mapper<Text, CopyListingFileStatus, Text, Text>.Context<Text, CopyListingFileStatus, Text, Text> context = stubContext.getContext();
        Configuration configuration = context.getConfiguration();
        configuration.setBoolean(DistCpOptionSwitch.IGNORE_FAILURES.getConfigLabel(), ignoreFailures);
        configuration.setBoolean(DistCpOptionSwitch.OVERWRITE.getConfigLabel(), true);
        configuration.setBoolean(DistCpOptionSwitch.SKIP_CRC.getConfigLabel(), true);
        copyMapper.setup(context);
        for (Path path : pathList) {
            final FileStatus fileStatus = fs.getFileStatus(path);
            if (!fileStatus.isDirectory()) {
                fs.delete(path, true);
                copyMapper.map(new Text(DistCpUtils.getRelativePath(new Path(SOURCE_PATH), path)), new CopyListingFileStatus(fileStatus), context);
            }
        }
        if (ignoreFailures) {
            for (Text value : stubContext.getWriter().values()) {
                Assert.assertTrue(value.toString() + " is not skipped", value.toString().startsWith("FAIL:"));
            }
        }
        Assert.assertTrue("There should have been an exception.", ignoreFailures);
    } catch (Exception e) {
        Assert.assertTrue("Unexpected exception: " + e.getMessage(), !ignoreFailures);
        e.printStackTrace();
    }
}
Also used : Path(org.apache.hadoop.fs.Path) Mapper(org.apache.hadoop.mapreduce.Mapper) FileStatus(org.apache.hadoop.fs.FileStatus) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) Configuration(org.apache.hadoop.conf.Configuration) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) FileSystem(org.apache.hadoop.fs.FileSystem) StubContext(org.apache.hadoop.tools.StubContext) Text(org.apache.hadoop.io.Text) IOException(java.io.IOException) AccessControlException(org.apache.hadoop.security.AccessControlException)

Example 38 with CopyListingFileStatus

use of org.apache.hadoop.tools.CopyListingFileStatus in project hadoop by apache.

the class TestCopyMapper method testPreserveBlockSizeAndReplicationImpl.

private void testPreserveBlockSizeAndReplicationImpl(boolean preserve) {
    try {
        deleteState();
        createSourceData();
        FileSystem fs = cluster.getFileSystem();
        CopyMapper copyMapper = new CopyMapper();
        StubContext stubContext = new StubContext(getConfiguration(), null, 0);
        Mapper<Text, CopyListingFileStatus, Text, Text>.Context<Text, CopyListingFileStatus, Text, Text> context = stubContext.getContext();
        Configuration configuration = context.getConfiguration();
        EnumSet<DistCpOptions.FileAttribute> fileAttributes = EnumSet.noneOf(DistCpOptions.FileAttribute.class);
        if (preserve) {
            fileAttributes.add(DistCpOptions.FileAttribute.BLOCKSIZE);
            fileAttributes.add(DistCpOptions.FileAttribute.REPLICATION);
        }
        configuration.set(DistCpOptionSwitch.PRESERVE_STATUS.getConfigLabel(), DistCpUtils.packAttributes(fileAttributes));
        copyMapper.setup(context);
        for (Path path : pathList) {
            final FileStatus fileStatus = fs.getFileStatus(path);
            copyMapper.map(new Text(DistCpUtils.getRelativePath(new Path(SOURCE_PATH), path)), new CopyListingFileStatus(fileStatus), context);
        }
        // Check that the block-size/replication aren't preserved.
        for (Path path : pathList) {
            final Path targetPath = new Path(path.toString().replaceAll(SOURCE_PATH, TARGET_PATH));
            final FileStatus source = fs.getFileStatus(path);
            final FileStatus target = fs.getFileStatus(targetPath);
            if (!source.isDirectory()) {
                Assert.assertTrue(preserve || source.getBlockSize() != target.getBlockSize());
                Assert.assertTrue(preserve || source.getReplication() != target.getReplication());
                Assert.assertTrue(!preserve || source.getBlockSize() == target.getBlockSize());
                Assert.assertTrue(!preserve || source.getReplication() == target.getReplication());
            }
        }
    } catch (Exception e) {
        Assert.assertTrue("Unexpected exception: " + e.getMessage(), false);
        e.printStackTrace();
    }
}
Also used : Path(org.apache.hadoop.fs.Path) FileStatus(org.apache.hadoop.fs.FileStatus) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) Configuration(org.apache.hadoop.conf.Configuration) Text(org.apache.hadoop.io.Text) IOException(java.io.IOException) AccessControlException(org.apache.hadoop.security.AccessControlException) Mapper(org.apache.hadoop.mapreduce.Mapper) DistCpOptions(org.apache.hadoop.tools.DistCpOptions) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) FileSystem(org.apache.hadoop.fs.FileSystem) StubContext(org.apache.hadoop.tools.StubContext)

Example 39 with CopyListingFileStatus

use of org.apache.hadoop.tools.CopyListingFileStatus in project hadoop by apache.

the class TestCopyMapper method testFileToDir.

@Test(timeout = 40000)
public void testFileToDir() {
    try {
        deleteState();
        createSourceData();
        FileSystem fs = cluster.getFileSystem();
        CopyMapper copyMapper = new CopyMapper();
        StubContext stubContext = new StubContext(getConfiguration(), null, 0);
        Mapper<Text, CopyListingFileStatus, Text, Text>.Context<Text, CopyListingFileStatus, Text, Text> context = stubContext.getContext();
        touchFile(SOURCE_PATH + "/src/file");
        mkdirs(TARGET_PATH + "/src/file");
        try {
            copyMapper.setup(context);
            copyMapper.map(new Text("/src/file"), new CopyListingFileStatus(fs.getFileStatus(new Path(SOURCE_PATH + "/src/file"))), context);
        } catch (IOException e) {
            Assert.assertTrue(e.getMessage().startsWith("Can't replace"));
        }
    } catch (Exception e) {
        LOG.error("Exception encountered ", e);
        Assert.fail("Test failed: " + e.getMessage());
    }
}
Also used : Path(org.apache.hadoop.fs.Path) Mapper(org.apache.hadoop.mapreduce.Mapper) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) FileSystem(org.apache.hadoop.fs.FileSystem) StubContext(org.apache.hadoop.tools.StubContext) Text(org.apache.hadoop.io.Text) IOException(java.io.IOException) IOException(java.io.IOException) AccessControlException(org.apache.hadoop.security.AccessControlException) Test(org.junit.Test)

Example 40 with CopyListingFileStatus

use of org.apache.hadoop.tools.CopyListingFileStatus in project hadoop by apache.

the class TestDistCpUtils method testPreserveTimestampOnDirectory.

@Test
public void testPreserveTimestampOnDirectory() throws IOException {
    FileSystem fs = FileSystem.get(config);
    EnumSet<FileAttribute> attributes = EnumSet.of(FileAttribute.TIMES);
    Path dst = new Path("/tmp/abc");
    Path src = new Path("/tmp/src");
    createDirectory(fs, src);
    createDirectory(fs, dst);
    fs.setPermission(src, fullPerm);
    fs.setOwner(src, "somebody", "somebody-group");
    fs.setTimes(src, 0, 0);
    fs.setPermission(dst, noPerm);
    fs.setOwner(dst, "nobody", "nobody-group");
    fs.setTimes(dst, 100, 100);
    CopyListingFileStatus srcStatus = new CopyListingFileStatus(fs.getFileStatus(src));
    DistCpUtils.preserve(fs, dst, srcStatus, attributes, false);
    CopyListingFileStatus dstStatus = new CopyListingFileStatus(fs.getFileStatus(dst));
    // FileStatus.equals only compares path field, must explicitly compare all fields
    Assert.assertFalse(srcStatus.getPermission().equals(dstStatus.getPermission()));
    Assert.assertFalse(srcStatus.getOwner().equals(dstStatus.getOwner()));
    Assert.assertFalse(srcStatus.getGroup().equals(dstStatus.getGroup()));
    Assert.assertTrue(srcStatus.getAccessTime() == dstStatus.getAccessTime());
    Assert.assertTrue(srcStatus.getModificationTime() == dstStatus.getModificationTime());
}
Also used : Path(org.apache.hadoop.fs.Path) CopyListingFileStatus(org.apache.hadoop.tools.CopyListingFileStatus) FileSystem(org.apache.hadoop.fs.FileSystem) FileAttribute(org.apache.hadoop.tools.DistCpOptions.FileAttribute) Test(org.junit.Test)

Aggregations

CopyListingFileStatus (org.apache.hadoop.tools.CopyListingFileStatus)44 Path (org.apache.hadoop.fs.Path)41 FileSystem (org.apache.hadoop.fs.FileSystem)36 Test (org.junit.Test)29 Text (org.apache.hadoop.io.Text)23 FileAttribute (org.apache.hadoop.tools.DistCpOptions.FileAttribute)20 StubContext (org.apache.hadoop.tools.StubContext)17 IOException (java.io.IOException)16 Mapper (org.apache.hadoop.mapreduce.Mapper)16 AccessControlException (org.apache.hadoop.security.AccessControlException)13 Configuration (org.apache.hadoop.conf.Configuration)11 DistCpOptions (org.apache.hadoop.tools.DistCpOptions)9 FileStatus (org.apache.hadoop.fs.FileStatus)8 FsPermission (org.apache.hadoop.fs.permission.FsPermission)6 SequenceFile (org.apache.hadoop.io.SequenceFile)5 UserGroupInformation (org.apache.hadoop.security.UserGroupInformation)5 ArrayList (java.util.ArrayList)3 OutputStream (java.io.OutputStream)2 AclEntry (org.apache.hadoop.fs.permission.AclEntry)2 FileSplit (org.apache.hadoop.mapreduce.lib.input.FileSplit)2