List of usage examples for org.apache.hadoop.fs FileStatus getModificationTime
public long getModificationTime()
From source file:org.apache.pig.impl.util.Utils.java
License:Apache License
/** * Finds a valid path for a file from an array of FileStatus objects. * @param statusArray Array of FileStatus objects in which to search * for the file.//from w ww. ja v a 2 s . co m * @param fileSystem FileSystem in which to search for the first file. * @return The first file found. * @throws IOException */ public static Path depthFirstSearchForFile(final FileStatus[] statusArray, final FileSystem fileSystem) throws IOException { // Most recent files first Arrays.sort(statusArray, new Comparator<FileStatus>() { @Override public int compare(final FileStatus fs1, final FileStatus fs2) { return Longs.compare(fs2.getModificationTime(), fs1.getModificationTime()); } }); for (FileStatus f : statusArray) { Path p = depthFirstSearchForFile(f, fileSystem); if (p != null) { return p; } } return null; }
From source file:org.apache.reef.runtime.yarn.client.uploader.JobFolder.java
License:Apache License
/** * Creates a LocalResource instance for the JAR file referenced by the given Path. *///from www .j a v a2 s . c o m public LocalResource getLocalResourceForPath(final Path jarPath) throws IOException { final LocalResource localResource = Records.newRecord(LocalResource.class); final FileStatus status = FileContext.getFileContext(fileSystem.getUri()).getFileStatus(jarPath); localResource.setType(LocalResourceType.ARCHIVE); localResource.setVisibility(LocalResourceVisibility.APPLICATION); localResource.setResource(ConverterUtils.getYarnUrlFromPath(status.getPath())); localResource.setTimestamp(status.getModificationTime()); localResource.setSize(status.getLen()); return localResource; }
From source file:org.apache.reef.runtime.yarn.client.YarnJobSubmissionHandler.java
License:Apache License
/** * Creates a LocalResource instance for the JAR file referenced by the given Path *//*from w w w. j ava2s .c o m*/ private LocalResource makeLocalResourceForJarFile(final Path path) throws IOException { final LocalResource localResource = Records.newRecord(LocalResource.class); final FileStatus status = FileContext.getFileContext(fileSystem.getUri()).getFileStatus(path); localResource.setType(LocalResourceType.ARCHIVE); localResource.setVisibility(LocalResourceVisibility.APPLICATION); localResource.setResource(ConverterUtils.getYarnUrlFromPath(status.getPath())); localResource.setTimestamp(status.getModificationTime()); localResource.setSize(status.getLen()); return localResource; }
From source file:org.apache.reef.runtime.yarn.driver.UploaderToJobFolder.java
License:Apache License
/** * Creates a LocalResource instance for the JAR file referenced by the given Path * * @param path/*from ww w. j av a 2 s. c o m*/ * @return * @throws IOException */ LocalResource makeLocalResourceForJarFile(final Path path) throws IOException { final LocalResource localResource = Records.newRecord(LocalResource.class); final FileStatus status = FileContext.getFileContext(this.fileSystem.getUri()).getFileStatus(path); localResource.setType(LocalResourceType.ARCHIVE); localResource.setVisibility(LocalResourceVisibility.APPLICATION); localResource.setResource(ConverterUtils.getYarnUrlFromPath(status.getPath())); localResource.setTimestamp(status.getModificationTime()); localResource.setSize(status.getLen()); return localResource; }
From source file:org.apache.samza.job.yarn.ContainerUtil.java
License:Apache License
protected void startContainer(Path packagePath, Container container, Map<String, String> env, final String cmd) { log.info("starting container {} {} {} {}", new Object[] { packagePath, container, env, cmd }); // set the local package so that the containers and app master are provisioned with it LocalResource packageResource = Records.newRecord(LocalResource.class); URL packageUrl = ConverterUtils.getYarnUrlFromPath(packagePath); FileStatus fileStatus; try {//from www. j a v a 2 s. co m fileStatus = packagePath.getFileSystem(yarnConfiguration).getFileStatus(packagePath); } catch (IOException ioe) { log.error("IO Exception when accessing the package status from the filesystem", ioe); throw new SamzaException("IO Exception when accessing the package status from the filesystem"); } packageResource.setResource(packageUrl); packageResource.setSize(fileStatus.getLen()); packageResource.setTimestamp(fileStatus.getModificationTime()); packageResource.setType(LocalResourceType.ARCHIVE); packageResource.setVisibility(LocalResourceVisibility.APPLICATION); ByteBuffer allTokens; // copy tokens (copied from dist shell example) try { Credentials credentials = UserGroupInformation.getCurrentUser().getCredentials(); DataOutputBuffer dob = new DataOutputBuffer(); credentials.writeTokenStorageToStream(dob); // now remove the AM->RM token so that containers cannot access it Iterator iter = credentials.getAllTokens().iterator(); while (iter.hasNext()) { TokenIdentifier token = ((Token) iter.next()).decodeIdentifier(); if (token.getKind().equals(AMRMTokenIdentifier.KIND_NAME)) { iter.remove(); } } allTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength()); } catch (IOException ioe) { ioe.printStackTrace(); throw new SamzaException("IO Exception when writing credentials to output buffer"); } ContainerLaunchContext context = Records.newRecord(ContainerLaunchContext.class); context.setEnvironment(env); context.setTokens(allTokens.duplicate()); context.setCommands(new ArrayList<String>() { { add(cmd); } }); context.setLocalResources(Collections.singletonMap("__package", packageResource)); log.debug("setting package to {}", packageResource); log.debug("setting context to {}", context); StartContainerRequest startContainerRequest = Records.newRecord(StartContainerRequest.class); startContainerRequest.setContainerLaunchContext(context); try { nmClient.startContainer(container, context); } catch (YarnException ye) { log.error("Received YarnException when starting container: " + container.getId(), ye); throw new SamzaException("Received YarnException when starting container: " + container.getId()); } catch (IOException ioe) { log.error("Received IOException when starting container: " + container.getId(), ioe); throw new SamzaException("Received IOException when starting container: " + container.getId()); } }
From source file:org.apache.samza.job.yarn.LocalizerResourceMapper.java
License:Apache License
private LocalResource createLocalResource(Path resourcePath, LocalResourceType resourceType, LocalResourceVisibility resourceVisibility) { LocalResource localResource = Records.newRecord(LocalResource.class); URL resourceUrl = ConverterUtils.getYarnUrlFromPath(resourcePath); try {//from w ww . j ava 2 s .c o m FileStatus resourceFileStatus = resourcePath.getFileSystem(yarnConfiguration) .getFileStatus(resourcePath); if (null == resourceFileStatus) { throw new LocalizerResourceException( "Check getFileStatus implementation. getFileStatus gets unexpected null for resourcePath " + resourcePath); } localResource.setResource(resourceUrl); log.info("setLocalizerResource for {}", resourceUrl); localResource.setSize(resourceFileStatus.getLen()); localResource.setTimestamp(resourceFileStatus.getModificationTime()); localResource.setType(resourceType); localResource.setVisibility(resourceVisibility); return localResource; } catch (IOException ioe) { log.error("IO Exception when accessing the resource file status from the filesystem: " + resourcePath, ioe); throw new LocalizerResourceException( "IO Exception when accessing the resource file status from the filesystem: " + resourcePath); } }
From source file:org.apache.samza.job.yarn.refactor.YarnContainerRunner.java
License:Apache License
/** * Runs a command as a process on the container. All binaries needed by the physical process are packaged in the URL * specified by packagePath./*from w w w . j a v a 2 s . c o m*/ */ private void startContainer(Path packagePath, Container container, Map<String, String> env, final String cmd) throws SamzaContainerLaunchException { log.info("starting container {} {} {} {}", new Object[] { packagePath, container, env, cmd }); // set the local package so that the containers and app master are provisioned with it LocalResource packageResource = Records.newRecord(LocalResource.class); URL packageUrl = ConverterUtils.getYarnUrlFromPath(packagePath); FileStatus fileStatus; try { fileStatus = packagePath.getFileSystem(yarnConfiguration).getFileStatus(packagePath); } catch (IOException ioe) { log.error("IO Exception when accessing the package status from the filesystem", ioe); throw new SamzaContainerLaunchException( "IO Exception when accessing the package status from the filesystem"); } packageResource.setResource(packageUrl); packageResource.setSize(fileStatus.getLen()); packageResource.setTimestamp(fileStatus.getModificationTime()); packageResource.setType(LocalResourceType.ARCHIVE); packageResource.setVisibility(LocalResourceVisibility.APPLICATION); ByteBuffer allTokens; // copy tokens (copied from dist shell example) try { Credentials credentials = UserGroupInformation.getCurrentUser().getCredentials(); DataOutputBuffer dob = new DataOutputBuffer(); credentials.writeTokenStorageToStream(dob); // now remove the AM->RM token so that containers cannot access it Iterator iter = credentials.getAllTokens().iterator(); while (iter.hasNext()) { TokenIdentifier token = ((Token) iter.next()).decodeIdentifier(); if (token.getKind().equals(AMRMTokenIdentifier.KIND_NAME)) { iter.remove(); } } allTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength()); } catch (IOException ioe) { log.error("IOException when writing credentials.", ioe); throw new SamzaContainerLaunchException("IO Exception when writing credentials to output buffer"); } ContainerLaunchContext context = Records.newRecord(ContainerLaunchContext.class); context.setEnvironment(env); context.setTokens(allTokens.duplicate()); context.setCommands(new ArrayList<String>() { { add(cmd); } }); context.setLocalResources(Collections.singletonMap("__package", packageResource)); log.debug("setting package to {}", packageResource); log.debug("setting context to {}", context); StartContainerRequest startContainerRequest = Records.newRecord(StartContainerRequest.class); startContainerRequest.setContainerLaunchContext(context); try { nmClient.startContainer(container, context); } catch (YarnException ye) { log.error("Received YarnException when starting container: " + container.getId(), ye); throw new SamzaContainerLaunchException( "Received YarnException when starting container: " + container.getId(), ye); } catch (IOException ioe) { log.error("Received IOException when starting container: " + container.getId(), ioe); throw new SamzaContainerLaunchException( "Received IOException when starting container: " + container.getId(), ioe); } }
From source file:org.apache.samza.job.yarn.YarnClusterResourceManager.java
License:Apache License
/** * Runs a command as a process on the container. All binaries needed by the physical process are packaged in the URL * specified by packagePath./*from w w w. j a va2 s .co m*/ */ private void startContainer(Path packagePath, Container container, Map<String, String> env, final String cmd) throws IOException { LocalResource packageResource = Records.newRecord(LocalResource.class); URL packageUrl = ConverterUtils.getYarnUrlFromPath(packagePath); FileStatus fileStatus; fileStatus = packagePath.getFileSystem(yarnConfiguration).getFileStatus(packagePath); packageResource.setResource(packageUrl); log.debug("Set package resource in YarnContainerRunner for {}", packageUrl); packageResource.setSize(fileStatus.getLen()); packageResource.setTimestamp(fileStatus.getModificationTime()); packageResource.setType(LocalResourceType.ARCHIVE); packageResource.setVisibility(LocalResourceVisibility.APPLICATION); ByteBuffer allTokens; // copy tokens to start the container Credentials credentials = UserGroupInformation.getCurrentUser().getCredentials(); DataOutputBuffer dob = new DataOutputBuffer(); credentials.writeTokenStorageToStream(dob); // now remove the AM->RM token so that containers cannot access it Iterator iter = credentials.getAllTokens().iterator(); while (iter.hasNext()) { TokenIdentifier token = ((org.apache.hadoop.security.token.Token) iter.next()).decodeIdentifier(); if (token != null && token.getKind().equals(AMRMTokenIdentifier.KIND_NAME)) { iter.remove(); } } allTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength()); Map<String, LocalResource> localResourceMap = new HashMap<>(); localResourceMap.put("__package", packageResource); // include the resources from the universal resource configurations LocalizerResourceMapper resourceMapper = new LocalizerResourceMapper(new LocalizerResourceConfig(config), yarnConfiguration); localResourceMap.putAll(resourceMapper.getResourceMap()); ContainerLaunchContext context = Records.newRecord(ContainerLaunchContext.class); context.setEnvironment(env); context.setTokens(allTokens.duplicate()); context.setCommands(new ArrayList<String>() { { add(cmd); } }); context.setLocalResources(localResourceMap); if (UserGroupInformation.isSecurityEnabled()) { Map<ApplicationAccessType, String> acls = yarnConfig.getYarnApplicationAcls(); if (!acls.isEmpty()) { context.setApplicationACLs(acls); } } log.debug("Setting localResourceMap to {}", localResourceMap); log.debug("Setting context to {}", context); StartContainerRequest startContainerRequest = Records.newRecord(StartContainerRequest.class); startContainerRequest.setContainerLaunchContext(context); log.info( "Making an async start request for Container ID: {} on host: {} with local resource map: {} and context: {}", container.getId(), container.getNodeHttpAddress(), localResourceMap.toString(), context); nmClientAsync.startContainerAsync(container, context); }
From source file:org.apache.samza.job.yarn.YarnContainerRunner.java
License:Apache License
/** * Runs a command as a process on the container. All binaries needed by the physical process are packaged in the URL * specified by packagePath./* w w w . j av a 2 s .com*/ */ private void startContainer(Path packagePath, Container container, Map<String, String> env, final String cmd) throws SamzaContainerLaunchException { log.info("starting container {} {} {} {}", new Object[] { packagePath, container, env, cmd }); // TODO: SAMZA-1144 remove the customized approach for package resource and use the common one. // But keep it now for backward compatibility. // set the local package so that the containers and app master are provisioned with it LocalResource packageResource = Records.newRecord(LocalResource.class); URL packageUrl = ConverterUtils.getYarnUrlFromPath(packagePath); FileStatus fileStatus; try { fileStatus = packagePath.getFileSystem(yarnConfiguration).getFileStatus(packagePath); } catch (IOException ioe) { log.error("IO Exception when accessing the package status from the filesystem", ioe); throw new SamzaContainerLaunchException( "IO Exception when accessing the package status from the filesystem"); } packageResource.setResource(packageUrl); log.info("set package Resource in YarnContainerRunner for {}", packageUrl); packageResource.setSize(fileStatus.getLen()); packageResource.setTimestamp(fileStatus.getModificationTime()); packageResource.setType(LocalResourceType.ARCHIVE); packageResource.setVisibility(LocalResourceVisibility.APPLICATION); ByteBuffer allTokens; // copy tokens (copied from dist shell example) try { Credentials credentials = UserGroupInformation.getCurrentUser().getCredentials(); DataOutputBuffer dob = new DataOutputBuffer(); credentials.writeTokenStorageToStream(dob); // now remove the AM->RM token so that containers cannot access it Iterator iter = credentials.getAllTokens().iterator(); while (iter.hasNext()) { TokenIdentifier token = ((Token) iter.next()).decodeIdentifier(); if (token != null && token.getKind().equals(AMRMTokenIdentifier.KIND_NAME)) { iter.remove(); } } allTokens = ByteBuffer.wrap(dob.getData(), 0, dob.getLength()); } catch (IOException ioe) { log.error("IOException when writing credentials.", ioe); throw new SamzaContainerLaunchException("IO Exception when writing credentials to output buffer"); } Map<String, LocalResource> localResourceMap = new HashMap<>(); localResourceMap.put("__package", packageResource); // include the resources from the universal resource configurations LocalizerResourceMapper resourceMapper = new LocalizerResourceMapper(new LocalizerResourceConfig(config), yarnConfiguration); localResourceMap.putAll(resourceMapper.getResourceMap()); ContainerLaunchContext context = Records.newRecord(ContainerLaunchContext.class); context.setEnvironment(env); context.setTokens(allTokens.duplicate()); context.setCommands(new ArrayList<String>() { { add(cmd); } }); context.setLocalResources(localResourceMap); log.debug("setting localResourceMap to {}", localResourceMap); log.debug("setting context to {}", context); StartContainerRequest startContainerRequest = Records.newRecord(StartContainerRequest.class); startContainerRequest.setContainerLaunchContext(context); try { nmClient.startContainer(container, context); } catch (YarnException ye) { log.error("Received YarnException when starting container: " + container.getId(), ye); throw new SamzaContainerLaunchException( "Received YarnException when starting container: " + container.getId(), ye); } catch (IOException ioe) { log.error("Received IOException when starting container: " + container.getId(), ioe); throw new SamzaContainerLaunchException( "Received IOException when starting container: " + container.getId(), ioe); } }
From source file:org.apache.solr.hadoop.morphline.MorphlineMapRunner.java
License:Apache License
protected Record getRecord(PathParts parts) { FileStatus stats; try {/*from w w w . jav a 2 s.co m*/ stats = parts.getFileStatus(); } catch (IOException e) { stats = null; } if (stats == null) { LOG.warn("Ignoring file that somehow has become unavailable since the job was submitted: {}", parts.getUploadURL()); return null; } Record headers = new Record(); //headers.put(getSchema().getUniqueKeyField().getName(), parts.getId()); // use HDFS file path as docId if no docId is specified headers.put(Fields.BASE_ID, parts.getId()); // with sanitizeUniqueKey command, use HDFS file path as docId if no docId is specified headers.put(Fields.ATTACHMENT_NAME, parts.getName()); // Tika can use the file name in guessing the right MIME type // enable indexing and storing of file meta data in Solr headers.put(HdfsFileFieldNames.FILE_UPLOAD_URL, parts.getUploadURL()); headers.put(HdfsFileFieldNames.FILE_DOWNLOAD_URL, parts.getDownloadURL()); headers.put(HdfsFileFieldNames.FILE_SCHEME, parts.getScheme()); headers.put(HdfsFileFieldNames.FILE_HOST, parts.getHost()); headers.put(HdfsFileFieldNames.FILE_PORT, String.valueOf(parts.getPort())); headers.put(HdfsFileFieldNames.FILE_PATH, parts.getURIPath()); headers.put(HdfsFileFieldNames.FILE_NAME, parts.getName()); headers.put(HdfsFileFieldNames.FILE_LAST_MODIFIED, String.valueOf(stats.getModificationTime())); // FIXME also add in SpoolDirectorySource headers.put(HdfsFileFieldNames.FILE_LENGTH, String.valueOf(stats.getLen())); // FIXME also add in SpoolDirectorySource headers.put(HdfsFileFieldNames.FILE_OWNER, stats.getOwner()); headers.put(HdfsFileFieldNames.FILE_GROUP, stats.getGroup()); headers.put(HdfsFileFieldNames.FILE_PERMISSIONS_USER, stats.getPermission().getUserAction().SYMBOL); headers.put(HdfsFileFieldNames.FILE_PERMISSIONS_GROUP, stats.getPermission().getGroupAction().SYMBOL); headers.put(HdfsFileFieldNames.FILE_PERMISSIONS_OTHER, stats.getPermission().getOtherAction().SYMBOL); headers.put(HdfsFileFieldNames.FILE_PERMISSIONS_STICKYBIT, String.valueOf(stats.getPermission().getStickyBit())); // TODO: consider to add stats.getAccessTime(), stats.getReplication(), stats.isSymlink(), stats.getBlockSize() return headers; }