[
https://issues.apache.org/jira/browse/STORM-876?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15014945#comment-15014945
]
ASF GitHub Bot commented on STORM-876:
--------------------------------------
Github user d2r commented on a diff in the pull request:
https://github.com/apache/storm/pull/845#discussion_r45425302
--- Diff:
external/storm-hdfs/src/main/java/org/apache/storm/hdfs/blobstore/HdfsBlobStore.java
---
@@ -0,0 +1,383 @@
+/**
+ * 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.storm.hdfs.blobstore;
+
+import backtype.storm.Config;
+import backtype.storm.blobstore.AtomicOutputStream;
+import backtype.storm.blobstore.AtomicOutputStream;
+import backtype.storm.blobstore.BlobStore;
+import backtype.storm.blobstore.BlobStoreAclHandler;
+import backtype.storm.blobstore.BlobStoreFile;
+import backtype.storm.blobstore.InputStreamWithMeta;
+import backtype.storm.generated.AuthorizationException;
+import backtype.storm.generated.KeyNotFoundException;
+import backtype.storm.generated.KeyAlreadyExistsException;
+import backtype.storm.generated.ReadableBlobMeta;
+import backtype.storm.generated.SettableBlobMeta;
+import backtype.storm.nimbus.NimbusInfo;
+import backtype.storm.utils.NimbusClient;
+import backtype.storm.utils.Utils;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.security.UserGroupInformation;
+import org.apache.thrift7.TBase;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.security.auth.Subject;
+import java.io.ByteArrayOutputStream;
+import java.io.FileNotFoundException;
+import java.io.IOException;
+import java.io.InputStream;
+import java.security.AccessController;
+import java.security.PrivilegedAction;
+import java.util.Iterator;
+import java.util.Map;
+
+import static backtype.storm.blobstore.BlobStoreAclHandler.ADMIN;
+import static backtype.storm.blobstore.BlobStoreAclHandler.READ;
+import static backtype.storm.blobstore.BlobStoreAclHandler.WRITE;
+
+/**
+ * Provides a HDFS file system backed blob store implementation.
+ * Note that this provides an api for having HDFS be the backing store for
the blobstore,
+ * it is not a service/daemon.
+ */
+public class HdfsBlobStore extends BlobStore {
+ public static final Logger LOG =
LoggerFactory.getLogger(HdfsBlobStore.class);
+ private static final String DATA_PREFIX = "data_";
+ private static final String META_PREFIX = "meta_";
+ private BlobStoreAclHandler _aclHandler;
+ private HdfsBlobStoreImpl _hbs;
+ private Subject _localSubject;
+ private Map conf;
+
+ /**
+ * Get the subject from Hadoop so we can use it to validate the acls.
There is no direct
+ * interface from UserGroupInformation to get the subject, so do a doAs
and get the context.
+ * We could probably run everything in the doAs but for now just grab
the subject.
+ */
+ private Subject getHadoopUser() {
+ Subject subj;
+ try {
+ subj = UserGroupInformation.getCurrentUser().doAs(
+ new PrivilegedAction<Subject>() {
+ @Override
+ public Subject run() {
+ return Subject.getSubject(AccessController.getContext());
+ }
+ });
+ } catch (IOException e) {
+ throw new RuntimeException("Error creating subject and logging user
in!", e);
+ }
+ return subj;
+ }
+
+ /**
+ * If who is null then we want to use the user hadoop says we are.
+ * Required for the supervisor to call these routines as its not
+ * logged in as anyone.
+ */
+ private Subject checkAndGetSubject(Subject who) {
+ if (who == null) {
+ return _localSubject;
+ }
+ return who;
+ }
+
+ @Override
+ public void prepare(Map conf, String overrideBase, NimbusInfo
nimbusInfo) {
+ this.conf = conf;
+ prepareInternal(conf, overrideBase, null);
+ }
+
+ /**
+ * Allow a Hadoop Configuration to be passed for testing. If it's null
then the hadoop configs
+ * must be in your classpath.
+ */
+ protected void prepareInternal(Map conf, String overrideBase,
Configuration hadoopConf) {
+ this.conf = conf;
+ if (overrideBase == null) {
+ overrideBase = (String)conf.get(Config.BLOBSTORE_DIR);
+ }
+ if (overrideBase == null) {
+ throw new RuntimeException("You must specify a blobstore directory
for HDFS to use!");
+ }
+ LOG.debug("directory is: {}", overrideBase);
+ try {
+ // if a HDFS keytab/principal have been supplied login, otherwise
assume they are
+ // logged in already or running insecure HDFS.
+ if (conf.get(Config.BLOBSTORE_HDFS_PRINCIPAL) != null &&
+ conf.get(Config.BLOBSTORE_HDFS_KEYTAB) != null) {
+ UserGroupInformation.loginUserFromKeytab((String)
conf.get(Config.BLOBSTORE_HDFS_PRINCIPAL),
+ (String) conf.get(Config.BLOBSTORE_HDFS_KEYTAB));
+ } else {
+ if (conf.get(Config.BLOBSTORE_HDFS_PRINCIPAL) == null &&
+ conf.get(Config.BLOBSTORE_HDFS_KEYTAB) != null) {
+ throw new RuntimeException("You must specify an HDFS principal
to go with the keytab!");
+
+ } else {
+ if (conf.get(Config.BLOBSTORE_HDFS_PRINCIPAL) != null &&
+ conf.get(Config.BLOBSTORE_HDFS_KEYTAB) == null) {
+ throw new RuntimeException("You must specify HDFS keytab go
with the principal!");
+ }
+ }
+ }
+ } catch (IOException e) {
+ throw new RuntimeException("Error logging in from keytab!", e);
+ }
+ Path baseDir = new Path(overrideBase, BASE_BLOBS_DIR_NAME);
+ try {
+ if (hadoopConf != null) {
+ _hbs = new HdfsBlobStoreImpl(baseDir, conf, hadoopConf);
+ } else {
+ _hbs = new HdfsBlobStoreImpl(baseDir, conf);
+ }
+ } catch (IOException e) {
+ throw new RuntimeException(e);
+ }
+ _localSubject = getHadoopUser();
+ _aclHandler = new BlobStoreAclHandler(conf);
+ }
+
+ @Override
+ public AtomicOutputStream createBlob(String key, SettableBlobMeta meta,
Subject who)
+ throws AuthorizationException, KeyAlreadyExistsException {
+ if (meta.get_replication_factor() <= 0) {
+
meta.set_replication_factor((int)conf.get(Config.STORM_BLOBSTORE_REPLICATION_FACTOR));
+ }
+ who = checkAndGetSubject(who);
+ validateKey(key);
+ _aclHandler.normalizeSettableBlobMeta(key, meta, who, READ | WRITE |
ADMIN);
+ BlobStoreAclHandler.validateSettableACLs(key, meta.get_acl());
+ _aclHandler.hasPermissions(meta.get_acl(), READ | WRITE | ADMIN, who,
key);
+ if (_hbs.exists(DATA_PREFIX+key)) {
+ throw new KeyAlreadyExistsException(key);
+ }
+ BlobStoreFileOutputStream mOut = null;
+ try {
+ BlobStoreFile metaFile = _hbs.write(META_PREFIX + key, true);
+ metaFile.setMetadata(meta);
+ mOut = new BlobStoreFileOutputStream(metaFile);
+ mOut.write(Utils.thriftSerialize((TBase) meta));
+ mOut.close();
+ mOut = null;
+ BlobStoreFile dataFile = _hbs.write(DATA_PREFIX + key, true);
+ dataFile.setMetadata(meta);
+ return new BlobStoreFileOutputStream(dataFile);
+ } catch (IOException e) {
+ throw new RuntimeException(e);
+ } finally {
+ if (mOut != null) {
+ try {
+ mOut.cancel();
+ } catch (IOException e) {
+ //Ignored
+ }
+ }
+ }
+ }
+
+ @Override
+ public AtomicOutputStream updateBlob(String key, Subject who)
+ throws AuthorizationException, KeyNotFoundException {
+ who = checkAndGetSubject(who);
+ SettableBlobMeta meta = getStoredBlobMeta(key);
+ validateKey(key);
+ _aclHandler.hasPermissions(meta.get_acl(), WRITE, who, key);
+ try {
+ BlobStoreFile dataFile = _hbs.write(DATA_PREFIX + key, false);
+ dataFile.setMetadata(meta);
+ return new BlobStoreFileOutputStream(dataFile);
+ } catch (IOException e) {
+ throw new RuntimeException(e);
+ }
+ }
+
+ private SettableBlobMeta getStoredBlobMeta(String key) throws
KeyNotFoundException {
+ InputStream in = null;
+ try {
+ BlobStoreFile pf = _hbs.read(META_PREFIX + key);
+ try {
+ in = pf.getInputStream();
+ } catch (FileNotFoundException fnf) {
+ throw new KeyNotFoundException(key);
+ }
+ ByteArrayOutputStream out = new ByteArrayOutputStream();
+ byte[] buffer = new byte[2048];
+ int len;
+ while ((len = in.read(buffer)) > 0) {
+ out.write(buffer, 0, len);
+ }
+ in.close();
+ in = null;
+ SettableBlobMeta sbm =
Utils.thriftDeserialize(SettableBlobMeta.class, out.toByteArray());
+ return sbm;
--- End diff --
We probably don't need to define `sbm` if we are just returning it here.
> Dist Cache: Basic Functionality
> -------------------------------
>
> Key: STORM-876
> URL: https://issues.apache.org/jira/browse/STORM-876
> Project: Apache Storm
> Issue Type: Improvement
> Components: storm-core
> Reporter: Robert Joseph Evans
> Assignee: Robert Joseph Evans
> Attachments: DISTCACHE.md, DistributedCacheDesignDocument.pdf
>
>
> Basic functionality for the Dist Cache feature.
> As part of this a new API should be added to support uploading and
> downloading dist cache items. storm-core.ser, storm-conf.ser and storm.jar
> should be written into the blob store instead of residing locally. We need a
> default implementation of the blob store that does essentially what nimbus
> currently does and does not need anything extra. But having an HDFS backend
> too would be great for scalability and HA.
> The supervisor should provide a way to download and manage these blobs and
> provide a working directory for the worker process with symlinks to the
> blobs. It should also allow the blobs to be updated and switch the symlink
> atomically to point to the new blob once it is downloaded.
> All of this is already done by code internal to Yahoo! we are in the process
> of getting it ready to push back to open source shortly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
