Search in sources :

Example 1 with TransferManager

use of com.amazonaws.services.s3.transfer.TransferManager in project alluxio by Alluxio.

the class S3AUnderFileSystem method createInstance.

/**
   * Constructs a new instance of {@link S3AUnderFileSystem}.
   *
   * @param uri the {@link AlluxioURI} for this UFS
   * @return the created {@link S3AUnderFileSystem} instance
   */
public static S3AUnderFileSystem createInstance(AlluxioURI uri) {
    String bucketName = uri.getHost();
    // Set the aws credential system properties based on Alluxio properties, if they are set
    if (Configuration.containsKey(PropertyKey.S3A_ACCESS_KEY)) {
        System.setProperty(SDKGlobalConfiguration.ACCESS_KEY_SYSTEM_PROPERTY, Configuration.get(PropertyKey.S3A_ACCESS_KEY));
    }
    if (Configuration.containsKey(PropertyKey.S3A_SECRET_KEY)) {
        System.setProperty(SDKGlobalConfiguration.SECRET_KEY_SYSTEM_PROPERTY, Configuration.get(PropertyKey.S3A_SECRET_KEY));
    }
    // Checks, in order, env variables, system properties, profile file, and instance profile
    AWSCredentialsProvider credentials = new AWSCredentialsProviderChain(new DefaultAWSCredentialsProviderChain());
    // Set the client configuration based on Alluxio configuration values
    ClientConfiguration clientConf = new ClientConfiguration();
    // Socket timeout
    clientConf.setSocketTimeout(Configuration.getInt(PropertyKey.UNDERFS_S3A_SOCKET_TIMEOUT_MS));
    // HTTP protocol
    if (Configuration.getBoolean(PropertyKey.UNDERFS_S3A_SECURE_HTTP_ENABLED)) {
        clientConf.setProtocol(Protocol.HTTPS);
    } else {
        clientConf.setProtocol(Protocol.HTTP);
    }
    // Proxy host
    if (Configuration.containsKey(PropertyKey.UNDERFS_S3_PROXY_HOST)) {
        clientConf.setProxyHost(Configuration.get(PropertyKey.UNDERFS_S3_PROXY_HOST));
    }
    // Proxy port
    if (Configuration.containsKey(PropertyKey.UNDERFS_S3_PROXY_PORT)) {
        clientConf.setProxyPort(Configuration.getInt(PropertyKey.UNDERFS_S3_PROXY_PORT));
    }
    int numAdminThreads = Configuration.getInt(PropertyKey.UNDERFS_S3_ADMIN_THREADS_MAX);
    int numTransferThreads = Configuration.getInt(PropertyKey.UNDERFS_S3_UPLOAD_THREADS_MAX);
    int numThreads = Configuration.getInt(PropertyKey.UNDERFS_S3_THREADS_MAX);
    if (numThreads < numAdminThreads + numTransferThreads) {
        LOG.warn("Configured s3 max threads: {} is less than # admin threads: {} plus transfer " + "threads {}. Using admin threads + transfer threads as max threads instead.");
        numThreads = numAdminThreads + numTransferThreads;
    }
    clientConf.setMaxConnections(numThreads);
    // Set client request timeout for all requests since multipart copy is used, and copy parts can
    // only be set with the client configuration.
    clientConf.setRequestTimeout(Configuration.getInt(PropertyKey.UNDERFS_S3A_REQUEST_TIMEOUT));
    AmazonS3Client amazonS3Client = new AmazonS3Client(credentials, clientConf);
    // Set a custom endpoint.
    if (Configuration.containsKey(PropertyKey.UNDERFS_S3_ENDPOINT)) {
        amazonS3Client.setEndpoint(Configuration.get(PropertyKey.UNDERFS_S3_ENDPOINT));
    }
    // Disable DNS style buckets, this enables path style requests.
    if (Configuration.getBoolean(PropertyKey.UNDERFS_S3_DISABLE_DNS_BUCKETS)) {
        S3ClientOptions clientOptions = S3ClientOptions.builder().setPathStyleAccess(true).build();
        amazonS3Client.setS3ClientOptions(clientOptions);
    }
    ExecutorService service = ExecutorServiceFactories.fixedThreadPoolExecutorServiceFactory("alluxio-s3-transfer-manager-worker", numTransferThreads).create();
    TransferManager transferManager = new TransferManager(amazonS3Client, service);
    TransferManagerConfiguration transferConf = new TransferManagerConfiguration();
    transferConf.setMultipartCopyThreshold(MULTIPART_COPY_THRESHOLD);
    transferManager.setConfiguration(transferConf);
    // Default to readable and writable by the user.
    short bucketMode = (short) 700;
    // There is no known account owner by default.
    String accountOwner = "";
    // if ACL enabled inherit bucket acl for all the objects.
    if (Configuration.getBoolean(PropertyKey.UNDERFS_S3A_INHERIT_ACL)) {
        String accountOwnerId = amazonS3Client.getS3AccountOwner().getId();
        // Gets the owner from user-defined static mapping from S3 canonical user
        // id to Alluxio user name.
        String owner = CommonUtils.getValueFromStaticMapping(Configuration.get(PropertyKey.UNDERFS_S3_OWNER_ID_TO_USERNAME_MAPPING), accountOwnerId);
        // If there is no user-defined mapping, use the display name.
        if (owner == null) {
            owner = amazonS3Client.getS3AccountOwner().getDisplayName();
        }
        accountOwner = owner == null ? accountOwnerId : owner;
        AccessControlList acl = amazonS3Client.getBucketAcl(bucketName);
        bucketMode = S3AUtils.translateBucketAcl(acl, accountOwnerId);
    }
    return new S3AUnderFileSystem(uri, amazonS3Client, bucketName, bucketMode, accountOwner, transferManager);
}
Also used : DefaultAWSCredentialsProviderChain(com.amazonaws.auth.DefaultAWSCredentialsProviderChain) AccessControlList(com.amazonaws.services.s3.model.AccessControlList) TransferManager(com.amazonaws.services.s3.transfer.TransferManager) AWSCredentialsProviderChain(com.amazonaws.auth.AWSCredentialsProviderChain) DefaultAWSCredentialsProviderChain(com.amazonaws.auth.DefaultAWSCredentialsProviderChain) AmazonS3Client(com.amazonaws.services.s3.AmazonS3Client) TransferManagerConfiguration(com.amazonaws.services.s3.transfer.TransferManagerConfiguration) S3ClientOptions(com.amazonaws.services.s3.S3ClientOptions) ExecutorService(java.util.concurrent.ExecutorService) AWSCredentialsProvider(com.amazonaws.auth.AWSCredentialsProvider) ClientConfiguration(com.amazonaws.ClientConfiguration)

Example 2 with TransferManager

use of com.amazonaws.services.s3.transfer.TransferManager in project jackrabbit-oak by apache.

the class S3Backend method init.

public void init(CachingDataStore store, String homeDir, Properties prop) throws DataStoreException {
    ClassLoader contextClassLoader = Thread.currentThread().getContextClassLoader();
    try {
        startTime = new Date();
        Thread.currentThread().setContextClassLoader(getClass().getClassLoader());
        LOG.debug("init");
        this.store = store;
        s3ReqDecorator = new S3RequestDecorator(prop);
        s3service = Utils.openService(prop);
        if (bucket == null || "".equals(bucket.trim())) {
            bucket = prop.getProperty(S3Constants.S3_BUCKET);
            // Alternately check if the 'container' property is set
            if (Strings.isNullOrEmpty(bucket)) {
                bucket = prop.getProperty(S3Constants.S3_CONTAINER);
            }
        }
        String region = prop.getProperty(S3Constants.S3_REGION);
        Region s3Region = null;
        if (StringUtils.isNullOrEmpty(region)) {
            com.amazonaws.regions.Region ec2Region = Regions.getCurrentRegion();
            if (ec2Region != null) {
                s3Region = Region.fromValue(ec2Region.getName());
            } else {
                throw new AmazonClientException("parameter [" + S3Constants.S3_REGION + "] not configured and cannot be derived from environment");
            }
        } else {
            if (Utils.DEFAULT_AWS_BUCKET_REGION.equals(region)) {
                s3Region = Region.US_Standard;
            } else if (Region.EU_Ireland.toString().equals(region)) {
                s3Region = Region.EU_Ireland;
            } else {
                s3Region = Region.fromValue(region);
            }
        }
        if (!s3service.doesBucketExist(bucket)) {
            s3service.createBucket(bucket, s3Region);
            LOG.info("Created bucket [{}] in [{}] ", bucket, region);
        } else {
            LOG.info("Using bucket [{}] in [{}] ", bucket, region);
        }
        int writeThreads = 10;
        String writeThreadsStr = prop.getProperty(S3Constants.S3_WRITE_THREADS);
        if (writeThreadsStr != null) {
            writeThreads = Integer.parseInt(writeThreadsStr);
        }
        LOG.info("Using thread pool of [{}] threads in S3 transfer manager.", writeThreads);
        tmx = new TransferManager(s3service, (ThreadPoolExecutor) Executors.newFixedThreadPool(writeThreads, new NamedThreadFactory("s3-transfer-manager-worker")));
        int asyncWritePoolSize = 10;
        String maxConnsStr = prop.getProperty(S3Constants.S3_MAX_CONNS);
        if (maxConnsStr != null) {
            asyncWritePoolSize = Integer.parseInt(maxConnsStr) - writeThreads;
        }
        asyncWriteExecuter = (ThreadPoolExecutor) Executors.newFixedThreadPool(asyncWritePoolSize, new NamedThreadFactory("s3-write-worker"));
        String renameKeyProp = prop.getProperty(S3Constants.S3_RENAME_KEYS);
        boolean renameKeyBool = (renameKeyProp == null || "".equals(renameKeyProp)) ? false : Boolean.parseBoolean(renameKeyProp);
        LOG.info("Rename keys [{}]", renameKeyBool);
        if (renameKeyBool) {
            renameKeys();
        }
        LOG.debug("S3 Backend initialized in [{}] ms", +(System.currentTimeMillis() - startTime.getTime()));
    } catch (Exception e) {
        LOG.debug("  error ", e);
        Map<String, String> filteredMap = Maps.newHashMap();
        if (prop != null) {
            filteredMap = Maps.filterKeys(Maps.fromProperties(prop), new Predicate<String>() {

                @Override
                public boolean apply(String input) {
                    return !input.equals(S3Constants.ACCESS_KEY) && !input.equals(S3Constants.SECRET_KEY);
                }
            });
        }
        throw new DataStoreException("Could not initialize S3 from " + filteredMap, e);
    } finally {
        if (contextClassLoader != null) {
            Thread.currentThread().setContextClassLoader(contextClassLoader);
        }
    }
}
Also used : TransferManager(com.amazonaws.services.s3.transfer.TransferManager) DataStoreException(org.apache.jackrabbit.core.data.DataStoreException) NamedThreadFactory(org.apache.jackrabbit.core.data.util.NamedThreadFactory) AmazonClientException(com.amazonaws.AmazonClientException) Date(java.util.Date) AmazonServiceException(com.amazonaws.AmazonServiceException) AmazonClientException(com.amazonaws.AmazonClientException) DataStoreException(org.apache.jackrabbit.core.data.DataStoreException) IOException(java.io.IOException) S3RequestDecorator(org.apache.jackrabbit.oak.blob.cloud.s3.S3RequestDecorator) Region(com.amazonaws.services.s3.model.Region) ThreadPoolExecutor(java.util.concurrent.ThreadPoolExecutor) Map(java.util.Map)

Example 3 with TransferManager

use of com.amazonaws.services.s3.transfer.TransferManager in project jackrabbit-oak by apache.

the class S3Backend method init.

public void init() throws DataStoreException {
    ClassLoader contextClassLoader = Thread.currentThread().getContextClassLoader();
    try {
        startTime = new Date();
        Thread.currentThread().setContextClassLoader(getClass().getClassLoader());
        LOG.debug("init");
        s3ReqDecorator = new S3RequestDecorator(properties);
        s3service = Utils.openService(properties);
        if (bucket == null || "".equals(bucket.trim())) {
            bucket = properties.getProperty(S3Constants.S3_BUCKET);
            // Alternately check if the 'container' property is set
            if (Strings.isNullOrEmpty(bucket)) {
                bucket = properties.getProperty(S3Constants.S3_CONTAINER);
            }
        }
        String region = properties.getProperty(S3Constants.S3_REGION);
        Region s3Region;
        if (StringUtils.isNullOrEmpty(region)) {
            com.amazonaws.regions.Region ec2Region = Regions.getCurrentRegion();
            if (ec2Region != null) {
                s3Region = Region.fromValue(ec2Region.getName());
            } else {
                throw new AmazonClientException("parameter [" + S3Constants.S3_REGION + "] not configured and cannot be derived from environment");
            }
        } else {
            if (Utils.DEFAULT_AWS_BUCKET_REGION.equals(region)) {
                s3Region = Region.US_Standard;
            } else if (Region.EU_Ireland.toString().equals(region)) {
                s3Region = Region.EU_Ireland;
            } else {
                s3Region = Region.fromValue(region);
            }
        }
        if (!s3service.doesBucketExist(bucket)) {
            s3service.createBucket(bucket, s3Region);
            LOG.info("Created bucket [{}] in [{}] ", bucket, region);
        } else {
            LOG.info("Using bucket [{}] in [{}] ", bucket, region);
        }
        int writeThreads = 10;
        String writeThreadsStr = properties.getProperty(S3Constants.S3_WRITE_THREADS);
        if (writeThreadsStr != null) {
            writeThreads = Integer.parseInt(writeThreadsStr);
        }
        LOG.info("Using thread pool of [{}] threads in S3 transfer manager.", writeThreads);
        tmx = new TransferManager(s3service, Executors.newFixedThreadPool(writeThreads, new NamedThreadFactory("s3-transfer-manager-worker")));
        String renameKeyProp = properties.getProperty(S3Constants.S3_RENAME_KEYS);
        boolean renameKeyBool = (renameKeyProp == null || "".equals(renameKeyProp)) ? false : Boolean.parseBoolean(renameKeyProp);
        LOG.info("Rename keys [{}]", renameKeyBool);
        if (renameKeyBool) {
            renameKeys();
        }
        LOG.debug("S3 Backend initialized in [{}] ms", +(System.currentTimeMillis() - startTime.getTime()));
    } catch (Exception e) {
        LOG.error("Error ", e);
        Map<String, Object> filteredMap = Maps.newHashMap();
        if (properties != null) {
            filteredMap = Maps.filterKeys(Utils.asMap(properties), new Predicate<String>() {

                @Override
                public boolean apply(String input) {
                    return !input.equals(S3Constants.ACCESS_KEY) && !input.equals(S3Constants.SECRET_KEY);
                }
            });
        }
        throw new DataStoreException("Could not initialize S3 from " + filteredMap, e);
    } finally {
        if (contextClassLoader != null) {
            Thread.currentThread().setContextClassLoader(contextClassLoader);
        }
    }
}
Also used : TransferManager(com.amazonaws.services.s3.transfer.TransferManager) DataStoreException(org.apache.jackrabbit.core.data.DataStoreException) NamedThreadFactory(org.apache.jackrabbit.core.data.util.NamedThreadFactory) AmazonClientException(com.amazonaws.AmazonClientException) Date(java.util.Date) AmazonServiceException(com.amazonaws.AmazonServiceException) AmazonClientException(com.amazonaws.AmazonClientException) DataStoreException(org.apache.jackrabbit.core.data.DataStoreException) IOException(java.io.IOException) Region(com.amazonaws.services.s3.model.Region) Map(java.util.Map)

Example 4 with TransferManager

use of com.amazonaws.services.s3.transfer.TransferManager in project herd by FINRAOS.

the class MockS3OperationsImpl method download.

@Override
public Download download(String bucket, String key, File file, TransferManager transferManager) {
    MockS3Bucket mockS3Bucket = mockS3Buckets.get(bucket);
    MockS3Object mockS3Object = mockS3Bucket.getObjects().get(key);
    try (FileOutputStream fileOutputStream = new FileOutputStream(file)) {
        fileOutputStream.write(mockS3Object.getData());
    } catch (IOException e) {
        throw new RuntimeException("Error writing to file " + file, e);
    }
    TransferProgress progress = new TransferProgress();
    progress.setTotalBytesToTransfer(mockS3Object.getData().length);
    progress.updateProgress(mockS3Object.getData().length);
    DownloadImpl download = new DownloadImpl(null, progress, null, null, null, new GetObjectRequest(bucket, key), file, mockS3Object.getObjectMetadata(), false);
    download.setState(TransferState.Completed);
    return download;
}
Also used : FileOutputStream(java.io.FileOutputStream) MultipleFileDownloadImpl(com.amazonaws.services.s3.transfer.internal.MultipleFileDownloadImpl) DownloadImpl(com.amazonaws.services.s3.transfer.internal.DownloadImpl) IOException(java.io.IOException) GetObjectRequest(com.amazonaws.services.s3.model.GetObjectRequest) TransferProgress(com.amazonaws.services.s3.transfer.TransferProgress)

Example 5 with TransferManager

use of com.amazonaws.services.s3.transfer.TransferManager in project herd by FINRAOS.

the class S3DaoImpl method downloadFile.

@Override
public S3FileTransferResultsDto downloadFile(final S3FileTransferRequestParamsDto params) throws InterruptedException {
    LOGGER.info("Downloading S3 file... s3Key=\"{}\" s3BucketName=\"{}\" localPath=\"{}\"", params.getS3KeyPrefix(), params.getS3BucketName(), params.getLocalPath());
    // Perform the transfer.
    S3FileTransferResultsDto results = performTransfer(params, new Transferer() {

        @Override
        public Transfer performTransfer(TransferManager transferManager) {
            return s3Operations.download(params.getS3BucketName(), params.getS3KeyPrefix(), new File(params.getLocalPath()), transferManager);
        }
    });
    LOGGER.info("Downloaded S3 file to the local system. s3Key=\"{}\" s3BucketName=\"{}\" localPath=\"{}\" totalBytesTransferred={} transferDuration=\"{}\"", params.getS3KeyPrefix(), params.getS3BucketName(), params.getLocalPath(), results.getTotalBytesTransferred(), HerdDateUtils.formatDuration(results.getDurationMillis()));
    logOverallTransferRate(results);
    return results;
}
Also used : TransferManager(com.amazonaws.services.s3.transfer.TransferManager) Transfer(com.amazonaws.services.s3.transfer.Transfer) S3FileTransferResultsDto(org.finra.herd.model.dto.S3FileTransferResultsDto) File(java.io.File)

Aggregations

TransferManager (com.amazonaws.services.s3.transfer.TransferManager)34 File (java.io.File)19 AmazonServiceException (com.amazonaws.AmazonServiceException)16 Upload (com.amazonaws.services.s3.transfer.Upload)10 AmazonClientException (com.amazonaws.AmazonClientException)8 ObjectMetadata (com.amazonaws.services.s3.model.ObjectMetadata)8 PutObjectRequest (com.amazonaws.services.s3.model.PutObjectRequest)8 Transfer (com.amazonaws.services.s3.transfer.Transfer)7 IOException (java.io.IOException)7 ArrayList (java.util.ArrayList)7 S3FileTransferResultsDto (org.finra.herd.model.dto.S3FileTransferResultsDto)7 AmazonS3 (com.amazonaws.services.s3.AmazonS3)6 Regions (com.amazonaws.regions.Regions)5 AmazonS3Client (com.amazonaws.services.s3.AmazonS3Client)5 MultipleFileUpload (com.amazonaws.services.s3.transfer.MultipleFileUpload)5 TransferManagerConfiguration (com.amazonaws.services.s3.transfer.TransferManagerConfiguration)5 TransferProgress (com.amazonaws.services.s3.transfer.TransferProgress)5 SdkClientException (com.amazonaws.SdkClientException)4 DefaultAWSCredentialsProviderChain (com.amazonaws.auth.DefaultAWSCredentialsProviderChain)4 ProfileCredentialsProvider (com.amazonaws.auth.profile.ProfileCredentialsProvider)4