HDFS-2197. Refactor RPC call implementations out of NameNode class. Contributed by Todd Lipcon.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/branches/branch-0.23@1169868 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Todd Lipcon 2011-09-12 18:59:42 +00:00
parent aae8027cd3
commit dacd3e5b18
42 changed files with 1226 additions and 1062 deletions

View File

@ -684,6 +684,7 @@ Release 0.23.0 - Unreleased
HDFS-1620. Rename HdfsConstants -> HdfsServerConstants, FSConstants -> HDFS-1620. Rename HdfsConstants -> HdfsServerConstants, FSConstants ->
HdfsConstants. (Harsh J Chouraria via atm) HdfsConstants. (Harsh J Chouraria via atm)
HDFS-2197. Refactor RPC call implementations out of NameNode class (todd)
OPTIMIZATIONS OPTIMIZATIONS

View File

@ -52,7 +52,7 @@ import org.apache.hadoop.net.NetUtils;
* </ol> * </ol>
*/ */
@InterfaceAudience.Private @InterfaceAudience.Private
public class BackupNode extends NameNode implements JournalProtocol { public class BackupNode extends NameNode {
private static final String BN_ADDRESS_NAME_KEY = DFSConfigKeys.DFS_NAMENODE_BACKUP_ADDRESS_KEY; private static final String BN_ADDRESS_NAME_KEY = DFSConfigKeys.DFS_NAMENODE_BACKUP_ADDRESS_KEY;
private static final String BN_ADDRESS_DEFAULT = DFSConfigKeys.DFS_NAMENODE_BACKUP_ADDRESS_DEFAULT; private static final String BN_ADDRESS_DEFAULT = DFSConfigKeys.DFS_NAMENODE_BACKUP_ADDRESS_DEFAULT;
private static final String BN_HTTP_ADDRESS_NAME_KEY = DFSConfigKeys.DFS_NAMENODE_BACKUP_HTTP_ADDRESS_KEY; private static final String BN_HTTP_ADDRESS_NAME_KEY = DFSConfigKeys.DFS_NAMENODE_BACKUP_HTTP_ADDRESS_KEY;
@ -95,18 +95,20 @@ public class BackupNode extends NameNode implements JournalProtocol {
} }
@Override // NameNode @Override // NameNode
protected void setRpcServerAddress(Configuration conf) { protected void setRpcServerAddress(Configuration conf,
conf.set(BN_ADDRESS_NAME_KEY, getHostPortString(rpcAddress)); InetSocketAddress addr) {
conf.set(BN_ADDRESS_NAME_KEY, getHostPortString(addr));
} }
@Override // Namenode @Override // Namenode
protected void setRpcServiceServerAddress(Configuration conf) { protected void setRpcServiceServerAddress(Configuration conf,
conf.set(BN_SERVICE_RPC_ADDRESS_KEY, getHostPortString(serviceRPCAddress)); InetSocketAddress addr) {
conf.set(BN_SERVICE_RPC_ADDRESS_KEY, getHostPortString(addr));
} }
@Override // NameNode @Override // NameNode
protected InetSocketAddress getHttpServerAddress(Configuration conf) { protected InetSocketAddress getHttpServerAddress(Configuration conf) {
assert rpcAddress != null : "rpcAddress should be calculated first"; assert getNameNodeAddress() != null : "rpcAddress should be calculated first";
String addr = conf.get(BN_HTTP_ADDRESS_NAME_KEY, BN_HTTP_ADDRESS_DEFAULT); String addr = conf.get(BN_HTTP_ADDRESS_NAME_KEY, BN_HTTP_ADDRESS_DEFAULT);
return NetUtils.createSocketAddr(addr); return NetUtils.createSocketAddr(addr);
} }
@ -145,6 +147,12 @@ public class BackupNode extends NameNode implements JournalProtocol {
runCheckpointDaemon(conf); runCheckpointDaemon(conf);
} }
@Override
protected NameNodeRpcServer createRpcServer(Configuration conf)
throws IOException {
return new BackupNodeRpcServer(conf, this);
}
@Override // NameNode @Override // NameNode
public void stop() { public void stop() {
if(checkpointManager != null) { if(checkpointManager != null) {
@ -177,75 +185,84 @@ public class BackupNode extends NameNode implements JournalProtocol {
super.stop(); super.stop();
} }
static class BackupNodeRpcServer extends NameNodeRpcServer implements JournalProtocol {
private final String nnRpcAddress;
@Override private BackupNodeRpcServer(Configuration conf, BackupNode nn)
public long getProtocolVersion(String protocol, long clientVersion) throws IOException {
throws IOException { super(conf, nn);
if (protocol.equals(JournalProtocol.class.getName())) { nnRpcAddress = nn.nnRpcAddress;
return JournalProtocol.versionID;
} else {
return super.getProtocolVersion(protocol, clientVersion);
} }
}
///////////////////////////////////////////////////// @Override
// NamenodeProtocol implementation for backup node. public long getProtocolVersion(String protocol, long clientVersion)
///////////////////////////////////////////////////// throws IOException {
@Override // NamenodeProtocol if (protocol.equals(JournalProtocol.class.getName())) {
public BlocksWithLocations getBlocks(DatanodeInfo datanode, long size) return JournalProtocol.versionID;
throws IOException { } else {
throw new UnsupportedActionException("getBlocks"); return super.getProtocolVersion(protocol, clientVersion);
} }
}
// Only active name-node can register other nodes. /////////////////////////////////////////////////////
@Override // NamenodeProtocol // NamenodeProtocol implementation for backup node.
public NamenodeRegistration register(NamenodeRegistration registration /////////////////////////////////////////////////////
) throws IOException { @Override // NamenodeProtocol
throw new UnsupportedActionException("register"); public BlocksWithLocations getBlocks(DatanodeInfo datanode, long size)
} throws IOException {
throw new UnsupportedActionException("getBlocks");
}
@Override // NamenodeProtocol // Only active name-node can register other nodes.
public NamenodeCommand startCheckpoint(NamenodeRegistration registration) @Override // NamenodeProtocol
throws IOException { public NamenodeRegistration register(NamenodeRegistration registration
throw new UnsupportedActionException("startCheckpoint"); ) throws IOException {
} throw new UnsupportedActionException("register");
}
@Override // NamenodeProtocol @Override // NamenodeProtocol
public void endCheckpoint(NamenodeRegistration registration, public NamenodeCommand startCheckpoint(NamenodeRegistration registration)
CheckpointSignature sig) throws IOException { throws IOException {
throw new UnsupportedActionException("endCheckpoint"); throw new UnsupportedActionException("startCheckpoint");
} }
///////////////////////////////////////////////////// @Override // NamenodeProtocol
// BackupNodeProtocol implementation for backup node. public void endCheckpoint(NamenodeRegistration registration,
///////////////////////////////////////////////////// CheckpointSignature sig) throws IOException {
throw new UnsupportedActionException("endCheckpoint");
}
@Override /////////////////////////////////////////////////////
public void journal(NamenodeRegistration nnReg, // BackupNodeProtocol implementation for backup node.
long firstTxId, int numTxns, /////////////////////////////////////////////////////
byte[] records) throws IOException {
verifyRequest(nnReg);
if(!nnRpcAddress.equals(nnReg.getAddress()))
throw new IOException("Journal request from unexpected name-node: "
+ nnReg.getAddress() + " expecting " + nnRpcAddress);
getBNImage().journal(firstTxId, numTxns, records);
}
@Override @Override
public void startLogSegment(NamenodeRegistration registration, long txid) public void journal(NamenodeRegistration nnReg,
throws IOException { long firstTxId, int numTxns,
verifyRequest(registration); byte[] records) throws IOException {
verifyRequest(nnReg);
if(!nnRpcAddress.equals(nnReg.getAddress()))
throw new IOException("Journal request from unexpected name-node: "
+ nnReg.getAddress() + " expecting " + rpcAddress);
getBNImage().journal(firstTxId, numTxns, records);
}
getBNImage().namenodeStartedLogSegment(txid); @Override
public void startLogSegment(NamenodeRegistration registration, long txid)
throws IOException {
verifyRequest(registration);
getBNImage().namenodeStartedLogSegment(txid);
}
private BackupImage getBNImage() {
return (BackupImage)nn.getFSImage();
}
} }
////////////////////////////////////////////////////// //////////////////////////////////////////////////////
BackupImage getBNImage() {
return (BackupImage)getFSImage();
}
boolean shouldCheckpointAtStartup() { boolean shouldCheckpointAtStartup() {
FSImage fsImage = getFSImage(); FSImage fsImage = getFSImage();
if(isRole(NamenodeRole.CHECKPOINT)) { if(isRole(NamenodeRole.CHECKPOINT)) {

View File

@ -69,7 +69,7 @@ public class CancelDelegationTokenServlet extends DfsServlet {
try { try {
ugi.doAs(new PrivilegedExceptionAction<Void>() { ugi.doAs(new PrivilegedExceptionAction<Void>() {
public Void run() throws Exception { public Void run() throws Exception {
nn.cancelDelegationToken(token); nn.getRpcServer().cancelDelegationToken(token);
return null; return null;
} }
}); });

View File

@ -73,7 +73,7 @@ abstract class DfsServlet extends HttpServlet {
// rpc // rpc
NameNode nn = NameNodeHttpServer.getNameNodeFromContext(context); NameNode nn = NameNodeHttpServer.getNameNodeFromContext(context);
if (nn != null) { if (nn != null) {
return nn; return nn.getRpcServer();
} }
InetSocketAddress nnAddr = InetSocketAddress nnAddr =
NameNodeHttpServer.getNameNodeAddressFromContext(context); NameNodeHttpServer.getNameNodeAddressFromContext(context);

View File

@ -75,7 +75,7 @@ public class GetDelegationTokenServlet extends DfsServlet {
+ ":" + NameNode.getAddress(conf).getPort(); + ":" + NameNode.getAddress(conf).getPort();
Token<DelegationTokenIdentifier> token = Token<DelegationTokenIdentifier> token =
nn.getDelegationToken(new Text(renewerFinal)); nn.getRpcServer().getDelegationToken(new Text(renewerFinal));
if(token == null) { if(token == null) {
throw new Exception("couldn't get the token for " +s); throw new Exception("couldn't get the token for " +s);
} }

View File

@ -0,0 +1,920 @@
/**
* 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.hadoop.hdfs.server.namenode;
import static org.apache.hadoop.hdfs.protocol.HdfsConstants.MAX_PATH_DEPTH;
import static org.apache.hadoop.hdfs.protocol.HdfsConstants.MAX_PATH_LENGTH;
import java.io.IOException;
import java.net.InetSocketAddress;
import java.util.Arrays;
import java.util.Collection;
import java.util.HashMap;
import org.apache.commons.logging.Log;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.CommonConfigurationKeys;
import org.apache.hadoop.fs.ContentSummary;
import org.apache.hadoop.fs.CreateFlag;
import org.apache.hadoop.fs.FsServerDefaults;
import org.apache.hadoop.fs.Options;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.UnresolvedLinkException;
import org.apache.hadoop.fs.permission.FsPermission;
import org.apache.hadoop.fs.permission.PermissionStatus;
import static org.apache.hadoop.hdfs.DFSConfigKeys.*;
import org.apache.hadoop.hdfs.HDFSPolicyProvider;
import org.apache.hadoop.hdfs.HdfsConfiguration;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.protocol.BlockListAsLongs;
import org.apache.hadoop.hdfs.protocol.ClientProtocol;
import org.apache.hadoop.hdfs.protocol.CorruptFileBlocks;
import org.apache.hadoop.hdfs.protocol.DatanodeID;
import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
import org.apache.hadoop.hdfs.protocol.DirectoryListing;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
import org.apache.hadoop.hdfs.protocol.LocatedBlock;
import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.protocol.UnregisteredNodeException;
import org.apache.hadoop.hdfs.protocol.UnresolvedPathException;
import org.apache.hadoop.hdfs.protocol.HdfsConstants.DatanodeReportType;
import org.apache.hadoop.hdfs.protocol.HdfsConstants.SafeModeAction;
import org.apache.hadoop.hdfs.protocol.HdfsConstants.UpgradeAction;
import org.apache.hadoop.hdfs.security.token.block.ExportedBlockKeys;
import org.apache.hadoop.hdfs.security.token.delegation.DelegationTokenIdentifier;
import org.apache.hadoop.hdfs.server.common.IncorrectVersionException;
import org.apache.hadoop.hdfs.server.common.UpgradeStatusReport;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.NamenodeRole;
import org.apache.hadoop.hdfs.server.namenode.metrics.NameNodeMetrics;
import org.apache.hadoop.hdfs.server.protocol.BlocksWithLocations;
import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand;
import org.apache.hadoop.hdfs.server.protocol.DatanodeProtocol;
import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.NamenodeCommand;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocol;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.server.protocol.NamenodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.NamespaceInfo;
import org.apache.hadoop.hdfs.server.protocol.NodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.RemoteEditLogManifest;
import org.apache.hadoop.hdfs.server.protocol.UpgradeCommand;
import org.apache.hadoop.io.EnumSetWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.ipc.ProtocolSignature;
import org.apache.hadoop.ipc.RPC;
import org.apache.hadoop.ipc.Server;
import org.apache.hadoop.net.Node;
import org.apache.hadoop.security.AccessControlException;
import org.apache.hadoop.security.Groups;
import org.apache.hadoop.security.RefreshUserMappingsProtocol;
import org.apache.hadoop.security.UserGroupInformation;
import org.apache.hadoop.security.authorize.AuthorizationException;
import org.apache.hadoop.security.authorize.ProxyUsers;
import org.apache.hadoop.security.authorize.RefreshAuthorizationPolicyProtocol;
import org.apache.hadoop.security.token.Token;
import org.apache.hadoop.security.token.SecretManager.InvalidToken;
import org.apache.hadoop.tools.GetUserMappingsProtocol;
/**
* This class is responsible for handling all of the RPC calls to the NameNode.
* It is created, started, and stopped by {@link NameNode}.
*/
class NameNodeRpcServer implements NamenodeProtocols {
private static final Log LOG = NameNode.LOG;
private static final Log stateChangeLog = NameNode.stateChangeLog;
// Dependencies from other parts of NN.
private final FSNamesystem namesystem;
protected final NameNode nn;
private final NameNodeMetrics metrics;
private final boolean serviceAuthEnabled;
/** The RPC server that listens to requests from DataNodes */
private final RPC.Server serviceRpcServer;
private final InetSocketAddress serviceRPCAddress;
/** The RPC server that listens to requests from clients */
protected final RPC.Server server;
protected final InetSocketAddress rpcAddress;
public NameNodeRpcServer(Configuration conf, NameNode nn)
throws IOException {
this.nn = nn;
this.namesystem = nn.getNamesystem();
this.metrics = NameNode.getNameNodeMetrics();
int handlerCount =
conf.getInt(DFS_DATANODE_HANDLER_COUNT_KEY,
DFS_DATANODE_HANDLER_COUNT_DEFAULT);
InetSocketAddress socAddr = nn.getRpcServerAddress(conf);
InetSocketAddress dnSocketAddr = nn.getServiceRpcServerAddress(conf);
if (dnSocketAddr != null) {
int serviceHandlerCount =
conf.getInt(DFS_NAMENODE_SERVICE_HANDLER_COUNT_KEY,
DFS_NAMENODE_SERVICE_HANDLER_COUNT_DEFAULT);
this.serviceRpcServer = RPC.getServer(NamenodeProtocols.class, this,
dnSocketAddr.getHostName(), dnSocketAddr.getPort(), serviceHandlerCount,
false, conf, namesystem.getDelegationTokenSecretManager());
this.serviceRPCAddress = this.serviceRpcServer.getListenerAddress();
nn.setRpcServiceServerAddress(conf, serviceRPCAddress);
} else {
serviceRpcServer = null;
serviceRPCAddress = null;
}
this.server = RPC.getServer(NamenodeProtocols.class, this,
socAddr.getHostName(), socAddr.getPort(),
handlerCount, false, conf,
namesystem.getDelegationTokenSecretManager());
// set service-level authorization security policy
if (serviceAuthEnabled =
conf.getBoolean(
CommonConfigurationKeys.HADOOP_SECURITY_AUTHORIZATION, false)) {
this.server.refreshServiceAcl(conf, new HDFSPolicyProvider());
if (this.serviceRpcServer != null) {
this.serviceRpcServer.refreshServiceAcl(conf, new HDFSPolicyProvider());
}
}
// The rpc-server port can be ephemeral... ensure we have the correct info
this.rpcAddress = this.server.getListenerAddress();
nn.setRpcServerAddress(conf, rpcAddress);
}
/**
* Actually start serving requests.
*/
void start() {
server.start(); //start RPC server
if (serviceRpcServer != null) {
serviceRpcServer.start();
}
}
/**
* Wait until the RPC server has shut down.
*/
void join() throws InterruptedException {
this.server.join();
}
void stop() {
if(server != null) server.stop();
if(serviceRpcServer != null) serviceRpcServer.stop();
}
InetSocketAddress getServiceRpcAddress() {
return serviceRPCAddress;
}
InetSocketAddress getRpcAddress() {
return rpcAddress;
}
@Override // VersionedProtocol
public ProtocolSignature getProtocolSignature(String protocol,
long clientVersion, int clientMethodsHash) throws IOException {
return ProtocolSignature.getProtocolSignature(
this, protocol, clientVersion, clientMethodsHash);
}
@Override
public long getProtocolVersion(String protocol,
long clientVersion) throws IOException {
if (protocol.equals(ClientProtocol.class.getName())) {
return ClientProtocol.versionID;
} else if (protocol.equals(DatanodeProtocol.class.getName())){
return DatanodeProtocol.versionID;
} else if (protocol.equals(NamenodeProtocol.class.getName())){
return NamenodeProtocol.versionID;
} else if (protocol.equals(RefreshAuthorizationPolicyProtocol.class.getName())){
return RefreshAuthorizationPolicyProtocol.versionID;
} else if (protocol.equals(RefreshUserMappingsProtocol.class.getName())){
return RefreshUserMappingsProtocol.versionID;
} else if (protocol.equals(GetUserMappingsProtocol.class.getName())){
return GetUserMappingsProtocol.versionID;
} else {
throw new IOException("Unknown protocol to name node: " + protocol);
}
}
/////////////////////////////////////////////////////
// NamenodeProtocol
/////////////////////////////////////////////////////
@Override // NamenodeProtocol
public BlocksWithLocations getBlocks(DatanodeInfo datanode, long size)
throws IOException {
if(size <= 0) {
throw new IllegalArgumentException(
"Unexpected not positive size: "+size);
}
return namesystem.getBlockManager().getBlocks(datanode, size);
}
@Override // NamenodeProtocol
public ExportedBlockKeys getBlockKeys() throws IOException {
return namesystem.getBlockManager().getBlockKeys();
}
@Override // NamenodeProtocol
public void errorReport(NamenodeRegistration registration,
int errorCode,
String msg) throws IOException {
verifyRequest(registration);
LOG.info("Error report from " + registration + ": " + msg);
if(errorCode == FATAL)
namesystem.releaseBackupNode(registration);
}
@Override // NamenodeProtocol
public NamenodeRegistration register(NamenodeRegistration registration)
throws IOException {
verifyVersion(registration.getVersion());
NamenodeRegistration myRegistration = nn.setRegistration();
namesystem.registerBackupNode(registration, myRegistration);
return myRegistration;
}
@Override // NamenodeProtocol
public NamenodeCommand startCheckpoint(NamenodeRegistration registration)
throws IOException {
verifyRequest(registration);
if(!nn.isRole(NamenodeRole.NAMENODE))
throw new IOException("Only an ACTIVE node can invoke startCheckpoint.");
return namesystem.startCheckpoint(registration, nn.setRegistration());
}
@Override // NamenodeProtocol
public void endCheckpoint(NamenodeRegistration registration,
CheckpointSignature sig) throws IOException {
verifyRequest(registration);
if(!nn.isRole(NamenodeRole.NAMENODE))
throw new IOException("Only an ACTIVE node can invoke endCheckpoint.");
namesystem.endCheckpoint(registration, sig);
}
@Override // ClientProtocol
public Token<DelegationTokenIdentifier> getDelegationToken(Text renewer)
throws IOException {
return namesystem.getDelegationToken(renewer);
}
@Override // ClientProtocol
public long renewDelegationToken(Token<DelegationTokenIdentifier> token)
throws InvalidToken, IOException {
return namesystem.renewDelegationToken(token);
}
@Override // ClientProtocol
public void cancelDelegationToken(Token<DelegationTokenIdentifier> token)
throws IOException {
namesystem.cancelDelegationToken(token);
}
@Override // ClientProtocol
public LocatedBlocks getBlockLocations(String src,
long offset,
long length)
throws IOException {
metrics.incrGetBlockLocations();
return namesystem.getBlockLocations(getClientMachine(),
src, offset, length);
}
@Override // ClientProtocol
public FsServerDefaults getServerDefaults() throws IOException {
return namesystem.getServerDefaults();
}
@Override // ClientProtocol
public void create(String src,
FsPermission masked,
String clientName,
EnumSetWritable<CreateFlag> flag,
boolean createParent,
short replication,
long blockSize) throws IOException {
String clientMachine = getClientMachine();
if (stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.create: file "
+src+" for "+clientName+" at "+clientMachine);
}
if (!checkPathLength(src)) {
throw new IOException("create: Pathname too long. Limit "
+ MAX_PATH_LENGTH + " characters, " + MAX_PATH_DEPTH + " levels.");
}
namesystem.startFile(src,
new PermissionStatus(UserGroupInformation.getCurrentUser().getShortUserName(),
null, masked),
clientName, clientMachine, flag.get(), createParent, replication, blockSize);
metrics.incrFilesCreated();
metrics.incrCreateFileOps();
}
@Override // ClientProtocol
public LocatedBlock append(String src, String clientName)
throws IOException {
String clientMachine = getClientMachine();
if (stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.append: file "
+src+" for "+clientName+" at "+clientMachine);
}
LocatedBlock info = namesystem.appendFile(src, clientName, clientMachine);
metrics.incrFilesAppended();
return info;
}
@Override // ClientProtocol
public boolean recoverLease(String src, String clientName) throws IOException {
String clientMachine = getClientMachine();
return namesystem.recoverLease(src, clientName, clientMachine);
}
@Override // ClientProtocol
public boolean setReplication(String src, short replication)
throws IOException {
return namesystem.setReplication(src, replication);
}
@Override // ClientProtocol
public void setPermission(String src, FsPermission permissions)
throws IOException {
namesystem.setPermission(src, permissions);
}
@Override // ClientProtocol
public void setOwner(String src, String username, String groupname)
throws IOException {
namesystem.setOwner(src, username, groupname);
}
@Override // ClientProtocol
public LocatedBlock addBlock(String src,
String clientName,
ExtendedBlock previous,
DatanodeInfo[] excludedNodes)
throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*BLOCK* NameNode.addBlock: file "
+src+" for "+clientName);
}
HashMap<Node, Node> excludedNodesSet = null;
if (excludedNodes != null) {
excludedNodesSet = new HashMap<Node, Node>(excludedNodes.length);
for (Node node:excludedNodes) {
excludedNodesSet.put(node, node);
}
}
LocatedBlock locatedBlock =
namesystem.getAdditionalBlock(src, clientName, previous, excludedNodesSet);
if (locatedBlock != null)
metrics.incrAddBlockOps();
return locatedBlock;
}
@Override // ClientProtocol
public LocatedBlock getAdditionalDatanode(final String src, final ExtendedBlock blk,
final DatanodeInfo[] existings, final DatanodeInfo[] excludes,
final int numAdditionalNodes, final String clientName
) throws IOException {
if (LOG.isDebugEnabled()) {
LOG.debug("getAdditionalDatanode: src=" + src
+ ", blk=" + blk
+ ", existings=" + Arrays.asList(existings)
+ ", excludes=" + Arrays.asList(excludes)
+ ", numAdditionalNodes=" + numAdditionalNodes
+ ", clientName=" + clientName);
}
metrics.incrGetAdditionalDatanodeOps();
HashMap<Node, Node> excludeSet = null;
if (excludes != null) {
excludeSet = new HashMap<Node, Node>(excludes.length);
for (Node node : excludes) {
excludeSet.put(node, node);
}
}
return namesystem.getAdditionalDatanode(src, blk,
existings, excludeSet, numAdditionalNodes, clientName);
}
/**
* The client needs to give up on the block.
*/
public void abandonBlock(ExtendedBlock b, String src, String holder)
throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*BLOCK* NameNode.abandonBlock: "
+b+" of file "+src);
}
if (!namesystem.abandonBlock(b, src, holder)) {
throw new IOException("Cannot abandon block during write to " + src);
}
}
@Override // ClientProtocol
public boolean complete(String src, String clientName, ExtendedBlock last)
throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.complete: "
+ src + " for " + clientName);
}
return namesystem.completeFile(src, clientName, last);
}
/**
* The client has detected an error on the specified located blocks
* and is reporting them to the server. For now, the namenode will
* mark the block as corrupt. In the future we might
* check the blocks are actually corrupt.
*/
@Override
public void reportBadBlocks(LocatedBlock[] blocks) throws IOException {
stateChangeLog.info("*DIR* NameNode.reportBadBlocks");
for (int i = 0; i < blocks.length; i++) {
ExtendedBlock blk = blocks[i].getBlock();
DatanodeInfo[] nodes = blocks[i].getLocations();
for (int j = 0; j < nodes.length; j++) {
DatanodeInfo dn = nodes[j];
namesystem.getBlockManager().findAndMarkBlockAsCorrupt(blk, dn);
}
}
}
@Override // ClientProtocol
public LocatedBlock updateBlockForPipeline(ExtendedBlock block, String clientName)
throws IOException {
return namesystem.updateBlockForPipeline(block, clientName);
}
@Override // ClientProtocol
public void updatePipeline(String clientName, ExtendedBlock oldBlock,
ExtendedBlock newBlock, DatanodeID[] newNodes)
throws IOException {
namesystem.updatePipeline(clientName, oldBlock, newBlock, newNodes);
}
@Override // DatanodeProtocol
public void commitBlockSynchronization(ExtendedBlock block,
long newgenerationstamp, long newlength,
boolean closeFile, boolean deleteblock, DatanodeID[] newtargets)
throws IOException {
namesystem.commitBlockSynchronization(block,
newgenerationstamp, newlength, closeFile, deleteblock, newtargets);
}
@Override // ClientProtocol
public long getPreferredBlockSize(String filename)
throws IOException {
return namesystem.getPreferredBlockSize(filename);
}
@Deprecated
@Override // ClientProtocol
public boolean rename(String src, String dst) throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.rename: " + src + " to " + dst);
}
if (!checkPathLength(dst)) {
throw new IOException("rename: Pathname too long. Limit "
+ MAX_PATH_LENGTH + " characters, " + MAX_PATH_DEPTH + " levels.");
}
boolean ret = namesystem.renameTo(src, dst);
if (ret) {
metrics.incrFilesRenamed();
}
return ret;
}
@Override // ClientProtocol
public void concat(String trg, String[] src) throws IOException {
namesystem.concat(trg, src);
}
@Override // ClientProtocol
public void rename(String src, String dst, Options.Rename... options)
throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.rename: " + src + " to " + dst);
}
if (!checkPathLength(dst)) {
throw new IOException("rename: Pathname too long. Limit "
+ MAX_PATH_LENGTH + " characters, " + MAX_PATH_DEPTH + " levels.");
}
namesystem.renameTo(src, dst, options);
metrics.incrFilesRenamed();
}
@Deprecated
@Override // ClientProtocol
public boolean delete(String src) throws IOException {
return delete(src, true);
}
@Override // ClientProtocol
public boolean delete(String src, boolean recursive) throws IOException {
if (stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* Namenode.delete: src=" + src
+ ", recursive=" + recursive);
}
boolean ret = namesystem.delete(src, recursive);
if (ret)
metrics.incrDeleteFileOps();
return ret;
}
/**
* Check path length does not exceed maximum. Returns true if
* length and depth are okay. Returns false if length is too long
* or depth is too great.
*/
private boolean checkPathLength(String src) {
Path srcPath = new Path(src);
return (src.length() <= MAX_PATH_LENGTH &&
srcPath.depth() <= MAX_PATH_DEPTH);
}
@Override // ClientProtocol
public boolean mkdirs(String src, FsPermission masked, boolean createParent)
throws IOException {
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*DIR* NameNode.mkdirs: " + src);
}
if (!checkPathLength(src)) {
throw new IOException("mkdirs: Pathname too long. Limit "
+ MAX_PATH_LENGTH + " characters, " + MAX_PATH_DEPTH + " levels.");
}
return namesystem.mkdirs(src,
new PermissionStatus(UserGroupInformation.getCurrentUser().getShortUserName(),
null, masked), createParent);
}
@Override // ClientProtocol
public void renewLease(String clientName) throws IOException {
namesystem.renewLease(clientName);
}
@Override // ClientProtocol
public DirectoryListing getListing(String src, byte[] startAfter,
boolean needLocation)
throws IOException {
DirectoryListing files = namesystem.getListing(
src, startAfter, needLocation);
if (files != null) {
metrics.incrGetListingOps();
metrics.incrFilesInGetListingOps(files.getPartialListing().length);
}
return files;
}
@Override // ClientProtocol
public HdfsFileStatus getFileInfo(String src) throws IOException {
metrics.incrFileInfoOps();
return namesystem.getFileInfo(src, true);
}
@Override // ClientProtocol
public HdfsFileStatus getFileLinkInfo(String src) throws IOException {
metrics.incrFileInfoOps();
return namesystem.getFileInfo(src, false);
}
@Override
public long[] getStats() {
return namesystem.getStats();
}
@Override // ClientProtocol
public DatanodeInfo[] getDatanodeReport(DatanodeReportType type)
throws IOException {
DatanodeInfo results[] = namesystem.datanodeReport(type);
if (results == null ) {
throw new IOException("Cannot find datanode report");
}
return results;
}
@Override // ClientProtocol
public boolean setSafeMode(SafeModeAction action) throws IOException {
return namesystem.setSafeMode(action);
}
@Override // ClientProtocol
public boolean restoreFailedStorage(String arg)
throws AccessControlException {
return namesystem.restoreFailedStorage(arg);
}
@Override // ClientProtocol
public void saveNamespace() throws IOException {
namesystem.saveNamespace();
}
@Override // ClientProtocol
public void refreshNodes() throws IOException {
namesystem.getBlockManager().getDatanodeManager().refreshNodes(
new HdfsConfiguration());
}
@Override // NamenodeProtocol
public long getTransactionID() {
return namesystem.getEditLog().getSyncTxId();
}
@Override // NamenodeProtocol
public CheckpointSignature rollEditLog() throws IOException {
return namesystem.rollEditLog();
}
@Override
public RemoteEditLogManifest getEditLogManifest(long sinceTxId)
throws IOException {
return namesystem.getEditLog().getEditLogManifest(sinceTxId);
}
@Override // ClientProtocol
public void finalizeUpgrade() throws IOException {
namesystem.finalizeUpgrade();
}
@Override // ClientProtocol
public UpgradeStatusReport distributedUpgradeProgress(UpgradeAction action)
throws IOException {
return namesystem.distributedUpgradeProgress(action);
}
@Override // ClientProtocol
public void metaSave(String filename) throws IOException {
namesystem.metaSave(filename);
}
@Override // ClientProtocol
public CorruptFileBlocks listCorruptFileBlocks(String path, String cookie)
throws IOException {
Collection<FSNamesystem.CorruptFileBlockInfo> fbs =
namesystem.listCorruptFileBlocks(path, cookie);
String[] files = new String[fbs.size()];
String lastCookie = "";
int i = 0;
for(FSNamesystem.CorruptFileBlockInfo fb: fbs) {
files[i++] = fb.path;
lastCookie = fb.block.getBlockName();
}
return new CorruptFileBlocks(files, lastCookie);
}
/**
* Tell all datanodes to use a new, non-persistent bandwidth value for
* dfs.datanode.balance.bandwidthPerSec.
* @param bandwidth Blanacer bandwidth in bytes per second for all datanodes.
* @throws IOException
*/
@Override // ClientProtocol
public void setBalancerBandwidth(long bandwidth) throws IOException {
namesystem.getBlockManager().getDatanodeManager().setBalancerBandwidth(bandwidth);
}
@Override // ClientProtocol
public ContentSummary getContentSummary(String path) throws IOException {
return namesystem.getContentSummary(path);
}
@Override // ClientProtocol
public void setQuota(String path, long namespaceQuota, long diskspaceQuota)
throws IOException {
namesystem.setQuota(path, namespaceQuota, diskspaceQuota);
}
@Override // ClientProtocol
public void fsync(String src, String clientName) throws IOException {
namesystem.fsync(src, clientName);
}
@Override // ClientProtocol
public void setTimes(String src, long mtime, long atime)
throws IOException {
namesystem.setTimes(src, mtime, atime);
}
@Override // ClientProtocol
public void createSymlink(String target, String link, FsPermission dirPerms,
boolean createParent) throws IOException {
metrics.incrCreateSymlinkOps();
/* We enforce the MAX_PATH_LENGTH limit even though a symlink target
* URI may refer to a non-HDFS file system.
*/
if (!checkPathLength(link)) {
throw new IOException("Symlink path exceeds " + MAX_PATH_LENGTH +
" character limit");
}
if ("".equals(target)) {
throw new IOException("Invalid symlink target");
}
final UserGroupInformation ugi = UserGroupInformation.getCurrentUser();
namesystem.createSymlink(target, link,
new PermissionStatus(ugi.getShortUserName(), null, dirPerms), createParent);
}
@Override // ClientProtocol
public String getLinkTarget(String path) throws IOException {
metrics.incrGetLinkTargetOps();
/* Resolves the first symlink in the given path, returning a
* new path consisting of the target of the symlink and any
* remaining path components from the original path.
*/
try {
HdfsFileStatus stat = namesystem.getFileInfo(path, false);
if (stat != null) {
// NB: getSymlink throws IOException if !stat.isSymlink()
return stat.getSymlink();
}
} catch (UnresolvedPathException e) {
return e.getResolvedPath().toString();
} catch (UnresolvedLinkException e) {
// The NameNode should only throw an UnresolvedPathException
throw new AssertionError("UnresolvedLinkException thrown");
}
return null;
}
@Override // DatanodeProtocol
public DatanodeRegistration registerDatanode(DatanodeRegistration nodeReg)
throws IOException {
verifyVersion(nodeReg.getVersion());
namesystem.registerDatanode(nodeReg);
return nodeReg;
}
@Override // DatanodeProtocol
public DatanodeCommand[] sendHeartbeat(DatanodeRegistration nodeReg,
long capacity, long dfsUsed, long remaining, long blockPoolUsed,
int xmitsInProgress, int xceiverCount, int failedVolumes)
throws IOException {
verifyRequest(nodeReg);
return namesystem.handleHeartbeat(nodeReg, capacity, dfsUsed, remaining,
blockPoolUsed, xceiverCount, xmitsInProgress, failedVolumes);
}
@Override // DatanodeProtocol
public DatanodeCommand blockReport(DatanodeRegistration nodeReg,
String poolId, long[] blocks) throws IOException {
verifyRequest(nodeReg);
BlockListAsLongs blist = new BlockListAsLongs(blocks);
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*BLOCK* NameNode.blockReport: "
+ "from " + nodeReg.getName() + " " + blist.getNumberOfBlocks()
+ " blocks");
}
namesystem.getBlockManager().processReport(nodeReg, poolId, blist);
if (nn.getFSImage().isUpgradeFinalized())
return new DatanodeCommand.Finalize(poolId);
return null;
}
@Override // DatanodeProtocol
public void blockReceived(DatanodeRegistration nodeReg, String poolId,
Block blocks[], String delHints[]) throws IOException {
verifyRequest(nodeReg);
if(stateChangeLog.isDebugEnabled()) {
stateChangeLog.debug("*BLOCK* NameNode.blockReceived: "
+"from "+nodeReg.getName()+" "+blocks.length+" blocks.");
}
for (int i = 0; i < blocks.length; i++) {
namesystem.getBlockManager().blockReceived(
nodeReg, poolId, blocks[i], delHints[i]);
}
}
@Override // DatanodeProtocol
public void errorReport(DatanodeRegistration nodeReg,
int errorCode, String msg) throws IOException {
String dnName = (nodeReg == null ? "unknown DataNode" : nodeReg.getName());
if (errorCode == DatanodeProtocol.NOTIFY) {
LOG.info("Error report from " + dnName + ": " + msg);
return;
}
verifyRequest(nodeReg);
if (errorCode == DatanodeProtocol.DISK_ERROR) {
LOG.warn("Disk error on " + dnName + ": " + msg);
} else if (errorCode == DatanodeProtocol.FATAL_DISK_ERROR) {
LOG.warn("Fatal disk error on " + dnName + ": " + msg);
namesystem.getBlockManager().getDatanodeManager().removeDatanode(nodeReg);
} else {
LOG.info("Error report from " + dnName + ": " + msg);
}
}
@Override // DatanodeProtocol, NamenodeProtocol
public NamespaceInfo versionRequest() throws IOException {
return namesystem.getNamespaceInfo();
}
@Override // DatanodeProtocol
public UpgradeCommand processUpgradeCommand(UpgradeCommand comm) throws IOException {
return namesystem.processDistributedUpgradeCommand(comm);
}
/**
* Verify request.
*
* Verifies correctness of the datanode version, registration ID, and
* if the datanode does not need to be shutdown.
*
* @param nodeReg data node registration
* @throws IOException
*/
void verifyRequest(NodeRegistration nodeReg) throws IOException {
verifyVersion(nodeReg.getVersion());
if (!namesystem.getRegistrationID().equals(nodeReg.getRegistrationID())) {
LOG.warn("Invalid registrationID - expected: "
+ namesystem.getRegistrationID() + " received: "
+ nodeReg.getRegistrationID());
throw new UnregisteredNodeException(nodeReg);
}
}
@Override // RefreshAuthorizationPolicyProtocol
public void refreshServiceAcl() throws IOException {
if (!serviceAuthEnabled) {
throw new AuthorizationException("Service Level Authorization not enabled!");
}
this.server.refreshServiceAcl(new Configuration(), new HDFSPolicyProvider());
if (this.serviceRpcServer != null) {
this.serviceRpcServer.refreshServiceAcl(new Configuration(), new HDFSPolicyProvider());
}
}
@Override // RefreshAuthorizationPolicyProtocol
public void refreshUserToGroupsMappings() throws IOException {
LOG.info("Refreshing all user-to-groups mappings. Requested by user: " +
UserGroupInformation.getCurrentUser().getShortUserName());
Groups.getUserToGroupsMappingService().refresh();
}
@Override // RefreshAuthorizationPolicyProtocol
public void refreshSuperUserGroupsConfiguration() {
LOG.info("Refreshing SuperUser proxy group mapping list ");
ProxyUsers.refreshSuperUserGroupsConfiguration();
}
@Override // GetUserMappingsProtocol
public String[] getGroupsForUser(String user) throws IOException {
if (LOG.isDebugEnabled()) {
LOG.debug("Getting groups for user " + user);
}
return UserGroupInformation.createRemoteUser(user).getGroupNames();
}
/**
* Verify version.
*
* @param version
* @throws IOException
*/
void verifyVersion(int version) throws IOException {
if (version != HdfsConstants.LAYOUT_VERSION)
throw new IncorrectVersionException(version, "data node");
}
private static String getClientMachine() {
String clientMachine = Server.getRemoteAddress();
if (clientMachine == null) {
clientMachine = "";
}
return clientMachine;
}
}

View File

@ -173,7 +173,7 @@ public class NamenodeFsck {
out.println(msg); out.println(msg);
namenode.getNamesystem().logFsckEvent(path, remoteAddress); namenode.getNamesystem().logFsckEvent(path, remoteAddress);
final HdfsFileStatus file = namenode.getFileInfo(path); final HdfsFileStatus file = namenode.getRpcServer().getFileInfo(path);
if (file != null) { if (file != null) {
if (showCorruptFileBlocks) { if (showCorruptFileBlocks) {
@ -250,7 +250,8 @@ public class NamenodeFsck {
res.totalDirs++; res.totalDirs++;
do { do {
assert lastReturnedName != null; assert lastReturnedName != null;
thisListing = namenode.getListing(path, lastReturnedName, false); thisListing = namenode.getRpcServer().getListing(
path, lastReturnedName, false);
if (thisListing == null) { if (thisListing == null) {
return; return;
} }
@ -385,7 +386,7 @@ public class NamenodeFsck {
break; break;
case FIXING_DELETE: case FIXING_DELETE:
if (!isOpen) if (!isOpen)
namenode.delete(path, true); namenode.getRpcServer().delete(path, true);
} }
} }
if (showFiles) { if (showFiles) {
@ -414,7 +415,8 @@ public class NamenodeFsck {
String target = lostFound + fullName; String target = lostFound + fullName;
String errmsg = "Failed to move " + fullName + " to /lost+found"; String errmsg = "Failed to move " + fullName + " to /lost+found";
try { try {
if (!namenode.mkdirs(target, file.getPermission(), true)) { if (!namenode.getRpcServer().mkdirs(
target, file.getPermission(), true)) {
LOG.warn(errmsg); LOG.warn(errmsg);
return; return;
} }

View File

@ -49,6 +49,7 @@ import org.apache.hadoop.hdfs.server.common.JspHelper;
import org.apache.hadoop.hdfs.server.common.Storage; import org.apache.hadoop.hdfs.server.common.Storage;
import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory; import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory;
import org.apache.hadoop.hdfs.server.common.UpgradeStatusReport; import org.apache.hadoop.hdfs.server.common.UpgradeStatusReport;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.io.Text; import org.apache.hadoop.io.Text;
import org.apache.hadoop.net.NodeBase; import org.apache.hadoop.net.NodeBase;
import org.apache.hadoop.security.UserGroupInformation; import org.apache.hadoop.security.UserGroupInformation;
@ -354,7 +355,7 @@ class NamenodeJspHelper {
} }
} }
static String getDelegationToken(final NameNode nn, static String getDelegationToken(final NamenodeProtocols nn,
HttpServletRequest request, Configuration conf, HttpServletRequest request, Configuration conf,
final UserGroupInformation ugi) throws IOException, InterruptedException { final UserGroupInformation ugi) throws IOException, InterruptedException {
Token<DelegationTokenIdentifier> token = ugi Token<DelegationTokenIdentifier> token = ugi
@ -381,7 +382,8 @@ class NamenodeJspHelper {
.getAttribute(JspHelper.CURRENT_CONF); .getAttribute(JspHelper.CURRENT_CONF);
final DatanodeID datanode = getRandomDatanode(nn); final DatanodeID datanode = getRandomDatanode(nn);
UserGroupInformation ugi = JspHelper.getUGI(context, request, conf); UserGroupInformation ugi = JspHelper.getUGI(context, request, conf);
String tokenString = getDelegationToken(nn, request, conf, ugi); String tokenString = getDelegationToken(
nn.getRpcServer(), request, conf, ugi);
// if the user is defined, get a delegation token and stringify it // if the user is defined, get a delegation token and stringify it
final String redirectLocation; final String redirectLocation;
final String nodeToRedirect; final String nodeToRedirect;

View File

@ -70,7 +70,7 @@ public class RenewDelegationTokenServlet extends DfsServlet {
try { try {
long result = ugi.doAs(new PrivilegedExceptionAction<Long>() { long result = ugi.doAs(new PrivilegedExceptionAction<Long>() {
public Long run() throws Exception { public Long run() throws Exception {
return nn.renewDelegationToken(token); return nn.getRpcServer().renewDelegationToken(token);
} }
}); });
PrintStream os = new PrintStream(resp.getOutputStream()); PrintStream os = new PrintStream(resp.getOutputStream());

View File

@ -1025,6 +1025,14 @@ public class MiniDFSCluster {
return getNameNode(0); return getNameNode(0);
} }
/**
* Get an instance of the NameNode's RPC handler.
*/
public NamenodeProtocols getNameNodeRpc() {
checkSingleNameNode();
return getNameNode(0).getRpcServer();
}
/** /**
* Gets the NameNode for the index. May be null. * Gets the NameNode for the index. May be null.
*/ */
@ -1361,7 +1369,15 @@ public class MiniDFSCluster {
if (nameNode == null) { if (nameNode == null) {
return false; return false;
} }
long[] sizes = nameNode.getStats(); long[] sizes;
try {
sizes = nameNode.getRpcServer().getStats();
} catch (IOException ioe) {
// This method above should never throw.
// It only throws IOE since it is exposed via RPC
throw new AssertionError("Unexpected IOE thrown: "
+ StringUtils.stringifyException(ioe));
}
boolean isUp = false; boolean isUp = false;
synchronized (this) { synchronized (this) {
isUp = ((!nameNode.isInSafeMode() || !waitSafeMode) && sizes[0] != 0); isUp = ((!nameNode.isInSafeMode() || !waitSafeMode) && sizes[0] != 0);

View File

@ -26,6 +26,7 @@ import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock; import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.server.namenode.LeaseExpiredException; import org.apache.hadoop.hdfs.server.namenode.LeaseExpiredException;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.HdfsConfiguration; import org.apache.hadoop.hdfs.HdfsConfiguration;
import org.apache.hadoop.io.IOUtils; import org.apache.hadoop.io.IOUtils;
@ -45,7 +46,7 @@ public class TestClientProtocolForPipelineRecovery {
try { try {
cluster.waitActive(); cluster.waitActive();
FileSystem fileSys = cluster.getFileSystem(); FileSystem fileSys = cluster.getFileSystem();
NameNode namenode = cluster.getNameNode(); NamenodeProtocols namenode = cluster.getNameNodeRpc();
/* Test writing to finalized replicas */ /* Test writing to finalized replicas */
Path file = new Path("dataprotocol.dat"); Path file = new Path("dataprotocol.dat");

View File

@ -57,6 +57,7 @@ import org.apache.hadoop.hdfs.protocol.LocatedBlock;
import org.apache.hadoop.hdfs.protocol.LocatedBlocks; import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.namenode.NotReplicatedYetException; import org.apache.hadoop.hdfs.server.namenode.NotReplicatedYetException;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.io.IOUtils; import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.ipc.RemoteException; import org.apache.hadoop.ipc.RemoteException;
import org.apache.hadoop.ipc.Client; import org.apache.hadoop.ipc.Client;
@ -190,7 +191,7 @@ public class TestDFSClientRetries extends TestCase {
final int maxRetries = 1; // Allow one retry (total of two calls) final int maxRetries = 1; // Allow one retry (total of two calls)
conf.setInt(DFSConfigKeys.DFS_CLIENT_BLOCK_WRITE_LOCATEFOLLOWINGBLOCK_RETRIES_KEY, maxRetries); conf.setInt(DFSConfigKeys.DFS_CLIENT_BLOCK_WRITE_LOCATEFOLLOWINGBLOCK_RETRIES_KEY, maxRetries);
NameNode mockNN = mock(NameNode.class); NamenodeProtocols mockNN = mock(NamenodeProtocols.class);
Answer<Object> answer = new ThrowsException(new IOException()) { Answer<Object> answer = new ThrowsException(new IOException()) {
int retryCount = 0; int retryCount = 0;
@ -240,8 +241,8 @@ public class TestDFSClientRetries extends TestCase {
try { try {
cluster.waitActive(); cluster.waitActive();
FileSystem fs = cluster.getFileSystem(); FileSystem fs = cluster.getFileSystem();
NameNode preSpyNN = cluster.getNameNode(); NamenodeProtocols preSpyNN = cluster.getNameNodeRpc();
NameNode spyNN = spy(preSpyNN); NamenodeProtocols spyNN = spy(preSpyNN);
DFSClient client = new DFSClient(null, spyNN, conf, null); DFSClient client = new DFSClient(null, spyNN, conf, null);
int maxBlockAcquires = client.getMaxBlockAcquireFailures(); int maxBlockAcquires = client.getMaxBlockAcquireFailures();
assertTrue(maxBlockAcquires > 0); assertTrue(maxBlockAcquires > 0);
@ -305,11 +306,11 @@ public class TestDFSClientRetries extends TestCase {
*/ */
private static class FailNTimesAnswer implements Answer<LocatedBlocks> { private static class FailNTimesAnswer implements Answer<LocatedBlocks> {
private int failuresLeft; private int failuresLeft;
private NameNode realNN; private NamenodeProtocols realNN;
public FailNTimesAnswer(NameNode realNN, int timesToFail) { public FailNTimesAnswer(NamenodeProtocols preSpyNN, int timesToFail) {
failuresLeft = timesToFail; failuresLeft = timesToFail;
this.realNN = realNN; this.realNN = preSpyNN;
} }
public LocatedBlocks answer(InvocationOnMock invocation) throws IOException { public LocatedBlocks answer(InvocationOnMock invocation) throws IOException {
@ -603,7 +604,8 @@ public class TestDFSClientRetries extends TestCase {
//stop the first datanode //stop the first datanode
final List<LocatedBlock> locatedblocks = DFSClient.callGetBlockLocations( final List<LocatedBlock> locatedblocks = DFSClient.callGetBlockLocations(
cluster.getNameNode(), f, 0, Long.MAX_VALUE).getLocatedBlocks(); cluster.getNameNodeRpc(), f, 0, Long.MAX_VALUE)
.getLocatedBlocks();
final DatanodeInfo first = locatedblocks.get(0).getLocations()[0]; final DatanodeInfo first = locatedblocks.get(0).getLocations()[0];
cluster.stopDataNode(first.getName()); cluster.stopDataNode(first.getName());

View File

@ -293,10 +293,11 @@ public class TestDecommission {
} }
private void verifyStats(NameNode namenode, FSNamesystem fsn, private void verifyStats(NameNode namenode, FSNamesystem fsn,
DatanodeInfo node, boolean decommissioning) throws InterruptedException { DatanodeInfo node, boolean decommissioning)
throws InterruptedException, IOException {
// Do the stats check over 10 iterations // Do the stats check over 10 iterations
for (int i = 0; i < 10; i++) { for (int i = 0; i < 10; i++) {
long[] newStats = namenode.getStats(); long[] newStats = namenode.getRpcServer().getStats();
// For decommissioning nodes, ensure capacity of the DN is no longer // For decommissioning nodes, ensure capacity of the DN is no longer
// counted. Only used space of the DN is counted in cluster capacity // counted. Only used space of the DN is counted in cluster capacity

View File

@ -36,7 +36,6 @@ import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataOutputStream; import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem; import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path; import org.apache.hadoop.fs.Path;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock; import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.protocol.HdfsConstants; import org.apache.hadoop.hdfs.protocol.HdfsConstants;
import org.apache.hadoop.hdfs.server.datanode.DataNode; import org.apache.hadoop.hdfs.server.datanode.DataNode;
@ -44,6 +43,7 @@ import org.apache.hadoop.hdfs.server.datanode.SimulatedFSDataset;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem; import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.LeaseManager; import org.apache.hadoop.hdfs.server.namenode.LeaseManager;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.test.GenericTestUtils; import org.apache.hadoop.test.GenericTestUtils;
import org.apache.log4j.Level; import org.apache.log4j.Level;
import org.junit.Before; import org.junit.Before;
@ -151,8 +151,8 @@ public class TestFileAppend4 {
try { try {
cluster.waitActive(); cluster.waitActive();
NameNode preSpyNN = cluster.getNameNode(); NamenodeProtocols preSpyNN = cluster.getNameNodeRpc();
NameNode spyNN = spy(preSpyNN); NamenodeProtocols spyNN = spy(preSpyNN);
// Delay completeFile // Delay completeFile
GenericTestUtils.DelayAnswer delayer = new GenericTestUtils.DelayAnswer(LOG); GenericTestUtils.DelayAnswer delayer = new GenericTestUtils.DelayAnswer(LOG);
@ -222,8 +222,8 @@ public class TestFileAppend4 {
try { try {
cluster.waitActive(); cluster.waitActive();
NameNode preSpyNN = cluster.getNameNode(); NamenodeProtocols preSpyNN = cluster.getNameNodeRpc();
NameNode spyNN = spy(preSpyNN); NamenodeProtocols spyNN = spy(preSpyNN);
// Delay completeFile // Delay completeFile
GenericTestUtils.DelayAnswer delayer = GenericTestUtils.DelayAnswer delayer =

View File

@ -420,7 +420,7 @@ public class TestFileCreation extends junit.framework.TestCase {
final Path f = new Path("/foo.txt"); final Path f = new Path("/foo.txt");
createFile(dfs, f, 3); createFile(dfs, f, 3);
try { try {
cluster.getNameNode().addBlock(f.toString(), cluster.getNameNodeRpc().addBlock(f.toString(),
client.clientName, null, null); client.clientName, null, null);
fail(); fail();
} catch(IOException ioe) { } catch(IOException ioe) {

View File

@ -106,7 +106,7 @@ public class TestLeaseRecovery extends junit.framework.TestCase {
DataNode.LOG.info("dfs.dfs.clientName=" + dfs.dfs.clientName); DataNode.LOG.info("dfs.dfs.clientName=" + dfs.dfs.clientName);
cluster.getNameNode().append(filestr, dfs.dfs.clientName); cluster.getNameNodeRpc().append(filestr, dfs.dfs.clientName);
// expire lease to trigger block recovery. // expire lease to trigger block recovery.
waitLeaseRecovery(cluster); waitLeaseRecovery(cluster);
@ -129,14 +129,14 @@ public class TestLeaseRecovery extends junit.framework.TestCase {
filestr = "/foo.safemode"; filestr = "/foo.safemode";
filepath = new Path(filestr); filepath = new Path(filestr);
dfs.create(filepath, (short)1); dfs.create(filepath, (short)1);
cluster.getNameNode().setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_ENTER); cluster.getNameNodeRpc().setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_ENTER);
assertTrue(dfs.dfs.exists(filestr)); assertTrue(dfs.dfs.exists(filestr));
DFSTestUtil.waitReplication(dfs, filepath, (short)1); DFSTestUtil.waitReplication(dfs, filepath, (short)1);
waitLeaseRecovery(cluster); waitLeaseRecovery(cluster);
// verify that we still cannot recover the lease // verify that we still cannot recover the lease
LeaseManager lm = NameNodeAdapter.getLeaseManager(cluster.getNamesystem()); LeaseManager lm = NameNodeAdapter.getLeaseManager(cluster.getNamesystem());
assertTrue("Found " + lm.countLease() + " lease, expected 1", lm.countLease() == 1); assertTrue("Found " + lm.countLease() + " lease, expected 1", lm.countLease() == 1);
cluster.getNameNode().setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); cluster.getNameNodeRpc().setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
} }
finally { finally {
if (cluster != null) {cluster.shutdown();} if (cluster != null) {cluster.shutdown();}

View File

@ -100,7 +100,7 @@ public class TestPipelines {
ofs.writeBytes("Some more stuff to write"); ofs.writeBytes("Some more stuff to write");
((DFSOutputStream) ofs.getWrappedStream()).hflush(); ((DFSOutputStream) ofs.getWrappedStream()).hflush();
List<LocatedBlock> lb = cluster.getNameNode().getBlockLocations( List<LocatedBlock> lb = cluster.getNameNodeRpc().getBlockLocations(
filePath.toString(), FILE_SIZE - 1, FILE_SIZE).getLocatedBlocks(); filePath.toString(), FILE_SIZE - 1, FILE_SIZE).getLocatedBlocks();
String bpid = cluster.getNamesystem().getBlockPoolId(); String bpid = cluster.getNamesystem().getBlockPoolId();

View File

@ -51,7 +51,7 @@ import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory;
import org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage; import org.apache.hadoop.hdfs.server.datanode.BlockPoolSliceStorage;
import org.apache.hadoop.hdfs.server.datanode.DataStorage; import org.apache.hadoop.hdfs.server.datanode.DataStorage;
import org.apache.hadoop.hdfs.server.namenode.NNStorage; import org.apache.hadoop.hdfs.server.namenode.NNStorage;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
/** /**
* This class defines a number of static helper methods used by the * This class defines a number of static helper methods used by the
@ -121,7 +121,7 @@ public class UpgradeUtilities {
.manageNameDfsDirs(false) .manageNameDfsDirs(false)
.build(); .build();
NameNode namenode = cluster.getNameNode(); NamenodeProtocols namenode = cluster.getNameNodeRpc();
namenodeStorageNamespaceID = namenode.versionRequest().getNamespaceID(); namenodeStorageNamespaceID = namenode.versionRequest().getNamespaceID();
namenodeStorageFsscTime = namenode.versionRequest().getCTime(); namenodeStorageFsscTime = namenode.versionRequest().getCTime();
namenodeStorageClusterID = namenode.versionRequest().getClusterID(); namenodeStorageClusterID = namenode.versionRequest().getClusterID();
@ -517,7 +517,7 @@ public class UpgradeUtilities {
*/ */
public static int getCurrentNamespaceID(MiniDFSCluster cluster) throws IOException { public static int getCurrentNamespaceID(MiniDFSCluster cluster) throws IOException {
if (cluster != null) { if (cluster != null) {
return cluster.getNameNode().versionRequest().getNamespaceID(); return cluster.getNameNodeRpc().versionRequest().getNamespaceID();
} }
return namenodeStorageNamespaceID; return namenodeStorageNamespaceID;
} }
@ -528,7 +528,7 @@ public class UpgradeUtilities {
*/ */
public static String getCurrentClusterID(MiniDFSCluster cluster) throws IOException { public static String getCurrentClusterID(MiniDFSCluster cluster) throws IOException {
if (cluster != null) { if (cluster != null) {
return cluster.getNameNode().versionRequest().getClusterID(); return cluster.getNameNodeRpc().versionRequest().getClusterID();
} }
return namenodeStorageClusterID; return namenodeStorageClusterID;
} }
@ -539,7 +539,7 @@ public class UpgradeUtilities {
*/ */
public static String getCurrentBlockPoolID(MiniDFSCluster cluster) throws IOException { public static String getCurrentBlockPoolID(MiniDFSCluster cluster) throws IOException {
if (cluster != null) { if (cluster != null) {
return cluster.getNameNode().versionRequest().getBlockPoolID(); return cluster.getNameNodeRpc().versionRequest().getBlockPoolID();
} }
return namenodeStorageBlockPoolID; return namenodeStorageBlockPoolID;
} }
@ -554,7 +554,7 @@ public class UpgradeUtilities {
*/ */
public static long getCurrentFsscTime(MiniDFSCluster cluster) throws IOException { public static long getCurrentFsscTime(MiniDFSCluster cluster) throws IOException {
if (cluster != null) { if (cluster != null) {
return cluster.getNameNode().versionRequest().getCTime(); return cluster.getNameNodeRpc().versionRequest().getCTime();
} }
return namenodeStorageFsscTime; return namenodeStorageFsscTime;
} }

View File

@ -375,11 +375,11 @@ public class TestBlockToken {
Path filePath = new Path(fileName); Path filePath = new Path(fileName);
FSDataOutputStream out = fs.create(filePath, (short) 1); FSDataOutputStream out = fs.create(filePath, (short) 1);
out.write(new byte[1000]); out.write(new byte[1000]);
LocatedBlocks locatedBlocks = cluster.getNameNode().getBlockLocations( LocatedBlocks locatedBlocks = cluster.getNameNodeRpc().getBlockLocations(
fileName, 0, 1000); fileName, 0, 1000);
while (locatedBlocks.getLastLocatedBlock() == null) { while (locatedBlocks.getLastLocatedBlock() == null) {
Thread.sleep(100); Thread.sleep(100);
locatedBlocks = cluster.getNameNode().getBlockLocations(fileName, 0, locatedBlocks = cluster.getNameNodeRpc().getBlockLocations(fileName, 0,
1000); 1000);
} }
Token<BlockTokenIdentifier> token = locatedBlocks.getLastLocatedBlock() Token<BlockTokenIdentifier> token = locatedBlocks.getLastLocatedBlock()

View File

@ -89,7 +89,7 @@ public class TestBalancerWithMultipleNameNodes {
this.cluster = cluster; this.cluster = cluster;
clients = new ClientProtocol[nNameNodes]; clients = new ClientProtocol[nNameNodes];
for(int i = 0; i < nNameNodes; i++) { for(int i = 0; i < nNameNodes; i++) {
clients[i] = cluster.getNameNode(i); clients[i] = cluster.getNameNode(i).getRpcServer();
} }
replication = (short)Math.max(1, nDataNodes - 1); replication = (short)Math.max(1, nDataNodes - 1);
} }

View File

@ -51,6 +51,7 @@ import org.apache.hadoop.hdfs.security.token.block.SecurityTestUtil;
import org.apache.hadoop.hdfs.server.balancer.TestBalancer; import org.apache.hadoop.hdfs.server.balancer.TestBalancer;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants; import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.net.NetUtils; import org.apache.hadoop.net.NetUtils;
import org.apache.hadoop.security.token.Token; import org.apache.hadoop.security.token.Token;
import org.apache.hadoop.util.StringUtils; import org.apache.hadoop.util.StringUtils;
@ -314,6 +315,7 @@ public class TestBlockTokenWithDFS {
assertEquals(numDataNodes, cluster.getDataNodes().size()); assertEquals(numDataNodes, cluster.getDataNodes().size());
final NameNode nn = cluster.getNameNode(); final NameNode nn = cluster.getNameNode();
final NamenodeProtocols nnProto = nn.getRpcServer();
final BlockManager bm = nn.getNamesystem().getBlockManager(); final BlockManager bm = nn.getNamesystem().getBlockManager();
final BlockTokenSecretManager sm = bm.getBlockTokenSecretManager(); final BlockTokenSecretManager sm = bm.getBlockTokenSecretManager();
@ -344,7 +346,7 @@ public class TestBlockTokenWithDFS {
new DFSClient(new InetSocketAddress("localhost", new DFSClient(new InetSocketAddress("localhost",
cluster.getNameNodePort()), conf); cluster.getNameNodePort()), conf);
List<LocatedBlock> locatedBlocks = nn.getBlockLocations( List<LocatedBlock> locatedBlocks = nnProto.getBlockLocations(
FILE_TO_READ, 0, FILE_SIZE).getLocatedBlocks(); FILE_TO_READ, 0, FILE_SIZE).getLocatedBlocks();
LocatedBlock lblock = locatedBlocks.get(0); // first block LocatedBlock lblock = locatedBlocks.get(0); // first block
Token<BlockTokenIdentifier> myToken = lblock.getBlockToken(); Token<BlockTokenIdentifier> myToken = lblock.getBlockToken();

View File

@ -34,7 +34,6 @@ import org.apache.hadoop.hdfs.server.blockmanagement.BlockManagerTestUtil;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants; import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem; import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.NameNode; import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.namenode.NameNodeAdapter;
import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand; import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand;
import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration; import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
import org.apache.hadoop.test.GenericTestUtils; import org.apache.hadoop.test.GenericTestUtils;
@ -139,7 +138,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N0); DataNode dn = cluster.getDataNodes().get(DN_N0);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
List<LocatedBlock> blocksAfterReport = List<LocatedBlock> blocksAfterReport =
@ -181,9 +180,10 @@ public class TestBlockReport {
List<ExtendedBlock> blocks2Remove = new ArrayList<ExtendedBlock>(); List<ExtendedBlock> blocks2Remove = new ArrayList<ExtendedBlock>();
List<Integer> removedIndex = new ArrayList<Integer>(); List<Integer> removedIndex = new ArrayList<Integer>();
List<LocatedBlock> lBlocks = cluster.getNameNode().getBlockLocations( List<LocatedBlock> lBlocks =
filePath.toString(), FILE_START, cluster.getNameNodeRpc().getBlockLocations(
FILE_SIZE).getLocatedBlocks(); filePath.toString(), FILE_START,
FILE_SIZE).getLocatedBlocks();
while (removedIndex.size() != 2) { while (removedIndex.size() != 2) {
int newRemoveIndex = rand.nextInt(lBlocks.size()); int newRemoveIndex = rand.nextInt(lBlocks.size());
@ -218,7 +218,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N0); DataNode dn = cluster.getDataNodes().get(DN_N0);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
BlockManagerTestUtil.getComputedDatanodeWork(cluster.getNamesystem() BlockManagerTestUtil.getComputedDatanodeWork(cluster.getNamesystem()
@ -258,7 +258,8 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N0); DataNode dn = cluster.getDataNodes().get(DN_N0);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
DatanodeCommand dnCmd = cluster.getNameNode().blockReport(dnR, poolId, DatanodeCommand dnCmd =
cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
if(LOG.isDebugEnabled()) { if(LOG.isDebugEnabled()) {
LOG.debug("Got the command: " + dnCmd); LOG.debug("Got the command: " + dnCmd);
@ -310,7 +311,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N1); DataNode dn = cluster.getDataNodes().get(DN_N1);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
printStats(); printStats();
assertEquals("Wrong number of PendingReplication Blocks", assertEquals("Wrong number of PendingReplication Blocks",
@ -359,7 +360,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N1); DataNode dn = cluster.getDataNodes().get(DN_N1);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
printStats(); printStats();
assertEquals("Wrong number of Corrupted blocks", assertEquals("Wrong number of Corrupted blocks",
@ -381,7 +382,7 @@ public class TestBlockReport {
LOG.debug("Done corrupting length of " + corruptedBlock.getBlockName()); LOG.debug("Done corrupting length of " + corruptedBlock.getBlockName());
} }
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
printStats(); printStats();
@ -431,7 +432,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N1); DataNode dn = cluster.getDataNodes().get(DN_N1);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
printStats(); printStats();
assertEquals("Wrong number of PendingReplication blocks", assertEquals("Wrong number of PendingReplication blocks",
@ -477,7 +478,7 @@ public class TestBlockReport {
DataNode dn = cluster.getDataNodes().get(DN_N1); DataNode dn = cluster.getDataNodes().get(DN_N1);
String poolId = cluster.getNamesystem().getBlockPoolId(); String poolId = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId); DatanodeRegistration dnR = dn.getDNRegistrationForBP(poolId);
cluster.getNameNode().blockReport(dnR, poolId, cluster.getNameNodeRpc().blockReport(dnR, poolId,
new BlockListAsLongs(blocks, null).getBlockListAsLongs()); new BlockListAsLongs(blocks, null).getBlockListAsLongs());
printStats(); printStats();
assertEquals("Wrong number of PendingReplication blocks", assertEquals("Wrong number of PendingReplication blocks",
@ -590,7 +591,7 @@ public class TestBlockReport {
DFSTestUtil.createFile(fs, filePath, fileSize, DFSTestUtil.createFile(fs, filePath, fileSize,
REPL_FACTOR, rand.nextLong()); REPL_FACTOR, rand.nextLong());
return locatedToBlocks(cluster.getNameNode() return locatedToBlocks(cluster.getNameNodeRpc()
.getBlockLocations(filePath.toString(), FILE_START, .getBlockLocations(filePath.toString(), FILE_START,
fileSize).getLocatedBlocks(), null); fileSize).getLocatedBlocks(), null);
} }
@ -707,7 +708,8 @@ public class TestBlockReport {
private Block findBlock(Path path, long size) throws IOException { private Block findBlock(Path path, long size) throws IOException {
Block ret; Block ret;
List<LocatedBlock> lbs = List<LocatedBlock> lbs =
cluster.getNameNode().getBlockLocations(path.toString(), cluster.getNameNodeRpc()
.getBlockLocations(path.toString(),
FILE_START, size).getLocatedBlocks(); FILE_START, size).getLocatedBlocks();
LocatedBlock lb = lbs.get(lbs.size() - 1); LocatedBlock lb = lbs.get(lbs.size() - 1);

View File

@ -40,8 +40,8 @@ import org.apache.hadoop.hdfs.protocol.LocatedBlock;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockManagerTestUtil; import org.apache.hadoop.hdfs.server.blockmanagement.BlockManagerTestUtil;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants; import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem; import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration; import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.DFSConfigKeys; import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.net.NetUtils; import org.apache.hadoop.net.NetUtils;
@ -144,7 +144,7 @@ public class TestDataNodeVolumeFailure {
String bpid = cluster.getNamesystem().getBlockPoolId(); String bpid = cluster.getNamesystem().getBlockPoolId();
DatanodeRegistration dnR = dn.getDNRegistrationForBP(bpid); DatanodeRegistration dnR = dn.getDNRegistrationForBP(bpid);
long[] bReport = dn.getFSDataset().getBlockReport(bpid).getBlockListAsLongs(); long[] bReport = dn.getFSDataset().getBlockReport(bpid).getBlockListAsLongs();
cluster.getNameNode().blockReport(dnR, bpid, bReport); cluster.getNameNodeRpc().blockReport(dnR, bpid, bReport);
// verify number of blocks and files... // verify number of blocks and files...
verify(filename, filesize); verify(filename, filesize);
@ -216,7 +216,7 @@ public class TestDataNodeVolumeFailure {
* @throws IOException * @throws IOException
*/ */
private void triggerFailure(String path, long size) throws IOException { private void triggerFailure(String path, long size) throws IOException {
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
List<LocatedBlock> locatedBlocks = List<LocatedBlock> locatedBlocks =
nn.getBlockLocations(path, 0, size).getLocatedBlocks(); nn.getBlockLocations(path, 0, size).getLocatedBlocks();
@ -291,7 +291,7 @@ public class TestDataNodeVolumeFailure {
throws IOException { throws IOException {
int total = 0; int total = 0;
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
List<LocatedBlock> locatedBlocks = List<LocatedBlock> locatedBlocks =
nn.getBlockLocations(path, 0, size).getLocatedBlocks(); nn.getBlockLocations(path, 0, size).getLocatedBlocks();
//System.out.println("Number of blocks: " + locatedBlocks.size()); //System.out.println("Number of blocks: " + locatedBlocks.size());

View File

@ -109,7 +109,7 @@ public class TestTransferRbw {
final DatanodeInfo oldnodeinfo; final DatanodeInfo oldnodeinfo;
{ {
final DatanodeInfo[] datatnodeinfos = cluster.getNameNode( final DatanodeInfo[] datatnodeinfos = cluster.getNameNodeRpc(
).getDatanodeReport(DatanodeReportType.LIVE); ).getDatanodeReport(DatanodeReportType.LIVE);
Assert.assertEquals(2, datatnodeinfos.length); Assert.assertEquals(2, datatnodeinfos.length);
int i = 0; int i = 0;

View File

@ -46,6 +46,7 @@ import org.apache.hadoop.hdfs.server.protocol.BlockCommand;
import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand; import org.apache.hadoop.hdfs.server.protocol.DatanodeCommand;
import org.apache.hadoop.hdfs.server.protocol.DatanodeProtocol; import org.apache.hadoop.hdfs.server.protocol.DatanodeProtocol;
import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration; import org.apache.hadoop.hdfs.server.protocol.DatanodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.server.protocol.NamespaceInfo; import org.apache.hadoop.hdfs.server.protocol.NamespaceInfo;
import org.apache.hadoop.hdfs.HdfsConfiguration; import org.apache.hadoop.hdfs.HdfsConfiguration;
import org.apache.hadoop.hdfs.DFSConfigKeys; import org.apache.hadoop.hdfs.DFSConfigKeys;
@ -100,6 +101,7 @@ public class NNThroughputBenchmark {
static Configuration config; static Configuration config;
static NameNode nameNode; static NameNode nameNode;
static NamenodeProtocols nameNodeProto;
NNThroughputBenchmark(Configuration conf) throws IOException, LoginException { NNThroughputBenchmark(Configuration conf) throws IOException, LoginException {
config = conf; config = conf;
@ -119,6 +121,7 @@ public class NNThroughputBenchmark {
// Start the NameNode // Start the NameNode
String[] argv = new String[] {}; String[] argv = new String[] {};
nameNode = NameNode.createNameNode(argv, config); nameNode = NameNode.createNameNode(argv, config);
nameNodeProto = nameNode.getRpcServer();
} }
void close() throws IOException { void close() throws IOException {
@ -264,9 +267,9 @@ public class NNThroughputBenchmark {
} }
void cleanUp() throws IOException { void cleanUp() throws IOException {
nameNode.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); nameNodeProto.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
if(!keepResults) if(!keepResults)
nameNode.delete(getBaseDir(), true); nameNodeProto.delete(getBaseDir(), true);
} }
int getNumOpsExecuted() { int getNumOpsExecuted() {
@ -397,7 +400,7 @@ public class NNThroughputBenchmark {
void benchmarkOne() throws IOException { void benchmarkOne() throws IOException {
for(int idx = 0; idx < opsPerThread; idx++) { for(int idx = 0; idx < opsPerThread; idx++) {
if((localNumOpsExecuted+1) % statsOp.ugcRefreshCount == 0) if((localNumOpsExecuted+1) % statsOp.ugcRefreshCount == 0)
nameNode.refreshUserToGroupsMappings(); nameNodeProto.refreshUserToGroupsMappings();
long stat = statsOp.executeOp(daemonId, idx, arg1); long stat = statsOp.executeOp(daemonId, idx, arg1);
localNumOpsExecuted++; localNumOpsExecuted++;
localCumulativeTime += stat; localCumulativeTime += stat;
@ -458,9 +461,9 @@ public class NNThroughputBenchmark {
*/ */
long executeOp(int daemonId, int inputIdx, String ignore) long executeOp(int daemonId, int inputIdx, String ignore)
throws IOException { throws IOException {
nameNode.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); nameNodeProto.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.delete(BASE_DIR_NAME, true); nameNodeProto.delete(BASE_DIR_NAME, true);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
} }
@ -522,7 +525,7 @@ public class NNThroughputBenchmark {
void generateInputs(int[] opsPerThread) throws IOException { void generateInputs(int[] opsPerThread) throws IOException {
assert opsPerThread.length == numThreads : "Error opsPerThread.length"; assert opsPerThread.length == numThreads : "Error opsPerThread.length";
nameNode.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); nameNodeProto.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
// int generatedFileIdx = 0; // int generatedFileIdx = 0;
LOG.info("Generate " + numOpsRequired + " intputs for " + getOpName()); LOG.info("Generate " + numOpsRequired + " intputs for " + getOpName());
fileNames = new String[numThreads][]; fileNames = new String[numThreads][];
@ -554,12 +557,12 @@ public class NNThroughputBenchmark {
throws IOException { throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
// dummyActionNoSynch(fileIdx); // dummyActionNoSynch(fileIdx);
nameNode.create(fileNames[daemonId][inputIdx], FsPermission.getDefault(), nameNodeProto.create(fileNames[daemonId][inputIdx], FsPermission.getDefault(),
clientName, new EnumSetWritable<CreateFlag>(EnumSet clientName, new EnumSetWritable<CreateFlag>(EnumSet
.of(CreateFlag.CREATE, CreateFlag.OVERWRITE)), true, replication, BLOCK_SIZE); .of(CreateFlag.CREATE, CreateFlag.OVERWRITE)), true, replication, BLOCK_SIZE);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
for(boolean written = !closeUponCreate; !written; for(boolean written = !closeUponCreate; !written;
written = nameNode.complete(fileNames[daemonId][inputIdx], written = nameNodeProto.complete(fileNames[daemonId][inputIdx],
clientName, null)); clientName, null));
return end-start; return end-start;
} }
@ -627,11 +630,11 @@ public class NNThroughputBenchmark {
} }
// use the same files for open // use the same files for open
super.generateInputs(opsPerThread); super.generateInputs(opsPerThread);
if(nameNode.getFileInfo(opCreate.getBaseDir()) != null if(nameNodeProto.getFileInfo(opCreate.getBaseDir()) != null
&& nameNode.getFileInfo(getBaseDir()) == null) { && nameNodeProto.getFileInfo(getBaseDir()) == null) {
nameNode.rename(opCreate.getBaseDir(), getBaseDir()); nameNodeProto.rename(opCreate.getBaseDir(), getBaseDir());
} }
if(nameNode.getFileInfo(getBaseDir()) == null) { if(nameNodeProto.getFileInfo(getBaseDir()) == null) {
throw new IOException(getBaseDir() + " does not exist."); throw new IOException(getBaseDir() + " does not exist.");
} }
} }
@ -642,7 +645,7 @@ public class NNThroughputBenchmark {
long executeOp(int daemonId, int inputIdx, String ignore) long executeOp(int daemonId, int inputIdx, String ignore)
throws IOException { throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.getBlockLocations(fileNames[daemonId][inputIdx], 0L, BLOCK_SIZE); nameNodeProto.getBlockLocations(fileNames[daemonId][inputIdx], 0L, BLOCK_SIZE);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
} }
@ -670,7 +673,7 @@ public class NNThroughputBenchmark {
long executeOp(int daemonId, int inputIdx, String ignore) long executeOp(int daemonId, int inputIdx, String ignore)
throws IOException { throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.delete(fileNames[daemonId][inputIdx], false); nameNodeProto.delete(fileNames[daemonId][inputIdx], false);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
} }
@ -698,7 +701,7 @@ public class NNThroughputBenchmark {
long executeOp(int daemonId, int inputIdx, String ignore) long executeOp(int daemonId, int inputIdx, String ignore)
throws IOException { throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.getFileInfo(fileNames[daemonId][inputIdx]); nameNodeProto.getFileInfo(fileNames[daemonId][inputIdx]);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
} }
@ -740,7 +743,7 @@ public class NNThroughputBenchmark {
long executeOp(int daemonId, int inputIdx, String ignore) long executeOp(int daemonId, int inputIdx, String ignore)
throws IOException { throws IOException {
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.rename(fileNames[daemonId][inputIdx], nameNodeProto.rename(fileNames[daemonId][inputIdx],
destNames[daemonId][inputIdx]); destNames[daemonId][inputIdx]);
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
@ -787,11 +790,11 @@ public class NNThroughputBenchmark {
void register() throws IOException { void register() throws IOException {
// get versions from the namenode // get versions from the namenode
nsInfo = nameNode.versionRequest(); nsInfo = nameNodeProto.versionRequest();
dnRegistration.setStorageInfo(new DataStorage(nsInfo, "")); dnRegistration.setStorageInfo(new DataStorage(nsInfo, ""));
DataNode.setNewStorageID(dnRegistration); DataNode.setNewStorageID(dnRegistration);
// register datanode // register datanode
dnRegistration = nameNode.registerDatanode(dnRegistration); dnRegistration = nameNodeProto.registerDatanode(dnRegistration);
} }
/** /**
@ -801,7 +804,7 @@ public class NNThroughputBenchmark {
void sendHeartbeat() throws IOException { void sendHeartbeat() throws IOException {
// register datanode // register datanode
// TODO:FEDERATION currently a single block pool is supported // TODO:FEDERATION currently a single block pool is supported
DatanodeCommand[] cmds = nameNode.sendHeartbeat(dnRegistration, DatanodeCommand[] cmds = nameNodeProto.sendHeartbeat(dnRegistration,
DF_CAPACITY, DF_USED, DF_CAPACITY - DF_USED, DF_USED, 0, 0, 0); DF_CAPACITY, DF_USED, DF_CAPACITY - DF_USED, DF_USED, 0, 0, 0);
if(cmds != null) { if(cmds != null) {
for (DatanodeCommand cmd : cmds ) { for (DatanodeCommand cmd : cmds ) {
@ -846,7 +849,7 @@ public class NNThroughputBenchmark {
int replicateBlocks() throws IOException { int replicateBlocks() throws IOException {
// register datanode // register datanode
// TODO:FEDERATION currently a single block pool is supported // TODO:FEDERATION currently a single block pool is supported
DatanodeCommand[] cmds = nameNode.sendHeartbeat(dnRegistration, DatanodeCommand[] cmds = nameNodeProto.sendHeartbeat(dnRegistration,
DF_CAPACITY, DF_USED, DF_CAPACITY - DF_USED, DF_USED, 0, 0, 0); DF_CAPACITY, DF_USED, DF_CAPACITY - DF_USED, DF_USED, 0, 0, 0);
if (cmds != null) { if (cmds != null) {
for (DatanodeCommand cmd : cmds) { for (DatanodeCommand cmd : cmds) {
@ -877,7 +880,7 @@ public class NNThroughputBenchmark {
receivedDNReg.setStorageInfo( receivedDNReg.setStorageInfo(
new DataStorage(nsInfo, dnInfo.getStorageID())); new DataStorage(nsInfo, dnInfo.getStorageID()));
receivedDNReg.setInfoPort(dnInfo.getInfoPort()); receivedDNReg.setInfoPort(dnInfo.getInfoPort());
nameNode.blockReceived( receivedDNReg, nameNodeProto.blockReceived( receivedDNReg,
nameNode.getNamesystem().getBlockPoolId(), nameNode.getNamesystem().getBlockPoolId(),
new Block[] {blocks[i]}, new Block[] {blocks[i]},
new String[] {DataNode.EMPTY_DEL_HINT}); new String[] {DataNode.EMPTY_DEL_HINT});
@ -968,14 +971,14 @@ public class NNThroughputBenchmark {
FileNameGenerator nameGenerator; FileNameGenerator nameGenerator;
nameGenerator = new FileNameGenerator(getBaseDir(), 100); nameGenerator = new FileNameGenerator(getBaseDir(), 100);
String clientName = getClientName(007); String clientName = getClientName(007);
nameNode.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); nameNodeProto.setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
for(int idx=0; idx < nrFiles; idx++) { for(int idx=0; idx < nrFiles; idx++) {
String fileName = nameGenerator.getNextFileName("ThroughputBench"); String fileName = nameGenerator.getNextFileName("ThroughputBench");
nameNode.create(fileName, FsPermission.getDefault(), clientName, nameNodeProto.create(fileName, FsPermission.getDefault(), clientName,
new EnumSetWritable<CreateFlag>(EnumSet.of(CreateFlag.CREATE, CreateFlag.OVERWRITE)), true, replication, new EnumSetWritable<CreateFlag>(EnumSet.of(CreateFlag.CREATE, CreateFlag.OVERWRITE)), true, replication,
BLOCK_SIZE); BLOCK_SIZE);
ExtendedBlock lastBlock = addBlocks(fileName, clientName); ExtendedBlock lastBlock = addBlocks(fileName, clientName);
nameNode.complete(fileName, clientName, lastBlock); nameNodeProto.complete(fileName, clientName, lastBlock);
} }
// prepare block reports // prepare block reports
for(int idx=0; idx < nrDatanodes; idx++) { for(int idx=0; idx < nrDatanodes; idx++) {
@ -987,12 +990,12 @@ public class NNThroughputBenchmark {
throws IOException { throws IOException {
ExtendedBlock prevBlock = null; ExtendedBlock prevBlock = null;
for(int jdx = 0; jdx < blocksPerFile; jdx++) { for(int jdx = 0; jdx < blocksPerFile; jdx++) {
LocatedBlock loc = nameNode.addBlock(fileName, clientName, prevBlock, null); LocatedBlock loc = nameNodeProto.addBlock(fileName, clientName, prevBlock, null);
prevBlock = loc.getBlock(); prevBlock = loc.getBlock();
for(DatanodeInfo dnInfo : loc.getLocations()) { for(DatanodeInfo dnInfo : loc.getLocations()) {
int dnIdx = Arrays.binarySearch(datanodes, dnInfo.getName()); int dnIdx = Arrays.binarySearch(datanodes, dnInfo.getName());
datanodes[dnIdx].addBlock(loc.getBlock().getLocalBlock()); datanodes[dnIdx].addBlock(loc.getBlock().getLocalBlock());
nameNode.blockReceived( nameNodeProto.blockReceived(
datanodes[dnIdx].dnRegistration, datanodes[dnIdx].dnRegistration,
loc.getBlock().getBlockPoolId(), loc.getBlock().getBlockPoolId(),
new Block[] {loc.getBlock().getLocalBlock()}, new Block[] {loc.getBlock().getLocalBlock()},
@ -1013,7 +1016,7 @@ public class NNThroughputBenchmark {
assert daemonId < numThreads : "Wrong daemonId."; assert daemonId < numThreads : "Wrong daemonId.";
TinyDatanode dn = datanodes[daemonId]; TinyDatanode dn = datanodes[daemonId];
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
nameNode.blockReport(dn.dnRegistration, nameNode.getNamesystem() nameNodeProto.blockReport(dn.dnRegistration, nameNode.getNamesystem()
.getBlockPoolId(), dn.getBlockReportList()); .getBlockPoolId(), dn.getBlockReportList());
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
return end-start; return end-start;
@ -1146,7 +1149,7 @@ public class NNThroughputBenchmark {
LOG.info("Datanode " + dn.getName() + " is decommissioned."); LOG.info("Datanode " + dn.getName() + " is decommissioned.");
} }
excludeFile.close(); excludeFile.close();
nameNode.refreshNodes(); nameNodeProto.refreshNodes();
} }
/** /**
@ -1160,8 +1163,8 @@ public class NNThroughputBenchmark {
assert daemonId < numThreads : "Wrong daemonId."; assert daemonId < numThreads : "Wrong daemonId.";
long start = System.currentTimeMillis(); long start = System.currentTimeMillis();
// compute data-node work // compute data-node work
int work = BlockManagerTestUtil.getComputedDatanodeWork(nameNode int work = BlockManagerTestUtil.getComputedDatanodeWork(
.getNamesystem().getBlockManager()); nameNode.getNamesystem().getBlockManager());
long end = System.currentTimeMillis(); long end = System.currentTimeMillis();
numPendingBlocks += work; numPendingBlocks += work;
if(work == 0) if(work == 0)

View File

@ -47,14 +47,6 @@ public class NameNodeAdapter {
src, offset, length, false, true); src, offset, length, false, true);
} }
/**
* Get the internal RPC server instance.
* @return rpc server
*/
public static Server getRpcServer(NameNode namenode) {
return namenode.server;
}
public static DelegationTokenSecretManager getDtSecretManager( public static DelegationTokenSecretManager getDtSecretManager(
final FSNamesystem ns) { final FSNamesystem ns) {
return ns.getDelegationTokenSecretManager(); return ns.getDelegationTokenSecretManager();

View File

@ -239,10 +239,10 @@ public class OfflineEditsViewerHelper {
LOG.info("Innocuous exception", e); LOG.info("Innocuous exception", e);
} }
locatedBlocks = DFSClientAdapter.callGetBlockLocations( locatedBlocks = DFSClientAdapter.callGetBlockLocations(
cluster.getNameNode(), filePath, 0L, bytes.length); cluster.getNameNodeRpc(), filePath, 0L, bytes.length);
} while (locatedBlocks.isUnderConstruction()); } while (locatedBlocks.isUnderConstruction());
// Force a roll so we get an OP_END_LOG_SEGMENT txn // Force a roll so we get an OP_END_LOG_SEGMENT txn
return cluster.getNameNode().rollEditLog(); return cluster.getNameNodeRpc().rollEditLog();
} }
} }

View File

@ -34,6 +34,7 @@ import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.StartupOption; import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.StartupOption;
import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory; import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory;
import org.apache.hadoop.hdfs.server.namenode.FileJournalManager.EditLogFile; import org.apache.hadoop.hdfs.server.namenode.FileJournalManager.EditLogFile;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.DFSConfigKeys; import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.test.GenericTestUtils; import org.apache.hadoop.test.GenericTestUtils;
import org.apache.log4j.Level; import org.apache.log4j.Level;
@ -128,12 +129,13 @@ public class TestBackupNode extends TestCase {
fileSys = cluster.getFileSystem(); fileSys = cluster.getFileSystem();
backup = startBackupNode(conf, StartupOption.BACKUP, 1); backup = startBackupNode(conf, StartupOption.BACKUP, 1);
BackupImage bnImage = backup.getBNImage(); BackupImage bnImage = (BackupImage) backup.getFSImage();
testBNInSync(cluster, backup, 1); testBNInSync(cluster, backup, 1);
// Force a roll -- BN should roll with NN. // Force a roll -- BN should roll with NN.
NameNode nn = cluster.getNameNode(); NameNode nn = cluster.getNameNode();
nn.rollEditLog(); NamenodeProtocols nnRpc = nn.getRpcServer();
nnRpc.rollEditLog();
assertEquals(bnImage.getEditLog().getCurSegmentTxId(), assertEquals(bnImage.getEditLog().getCurSegmentTxId(),
nn.getFSImage().getEditLog().getCurSegmentTxId()); nn.getFSImage().getEditLog().getCurSegmentTxId());
@ -207,7 +209,9 @@ public class TestBackupNode extends TestCase {
LOG.info("Checking for " + src + " on BN"); LOG.info("Checking for " + src + " on BN");
try { try {
boolean hasFile = backup.getNamesystem().getFileInfo(src, false) != null; boolean hasFile = backup.getNamesystem().getFileInfo(src, false) != null;
boolean txnIdMatch = backup.getTransactionID() == nn.getTransactionID(); boolean txnIdMatch =
backup.getRpcServer().getTransactionID() ==
nn.getRpcServer().getTransactionID();
return hasFile && txnIdMatch; return hasFile && txnIdMatch;
} catch (Exception e) { } catch (Exception e) {
throw new RuntimeException(e); throw new RuntimeException(e);
@ -264,7 +268,7 @@ public class TestBackupNode extends TestCase {
// //
// Take a checkpoint // Take a checkpoint
// //
long txid = cluster.getNameNode().getTransactionID(); long txid = cluster.getNameNodeRpc().getTransactionID();
backup = startBackupNode(conf, op, 1); backup = startBackupNode(conf, op, 1);
waitCheckpointDone(cluster, backup, txid); waitCheckpointDone(cluster, backup, txid);
} catch(IOException e) { } catch(IOException e) {
@ -300,18 +304,18 @@ public class TestBackupNode extends TestCase {
// Take a checkpoint // Take a checkpoint
// //
backup = startBackupNode(conf, op, 1); backup = startBackupNode(conf, op, 1);
long txid = cluster.getNameNode().getTransactionID(); long txid = cluster.getNameNodeRpc().getTransactionID();
waitCheckpointDone(cluster, backup, txid); waitCheckpointDone(cluster, backup, txid);
for (int i = 0; i < 10; i++) { for (int i = 0; i < 10; i++) {
fileSys.mkdirs(new Path("file_" + i)); fileSys.mkdirs(new Path("file_" + i));
} }
txid = cluster.getNameNode().getTransactionID(); txid = cluster.getNameNodeRpc().getTransactionID();
backup.doCheckpoint(); backup.doCheckpoint();
waitCheckpointDone(cluster, backup, txid); waitCheckpointDone(cluster, backup, txid);
txid = cluster.getNameNode().getTransactionID(); txid = cluster.getNameNodeRpc().getTransactionID();
backup.doCheckpoint(); backup.doCheckpoint();
waitCheckpointDone(cluster, backup, txid); waitCheckpointDone(cluster, backup, txid);

View File

@ -38,6 +38,7 @@ import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo; import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoUnderConstruction; import org.apache.hadoop.hdfs.server.blockmanagement.BlockInfoUnderConstruction;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.BlockUCState; import org.apache.hadoop.hdfs.server.common.HdfsServerConstants.BlockUCState;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.junit.AfterClass; import org.junit.AfterClass;
import org.junit.BeforeClass; import org.junit.BeforeClass;
import org.junit.Test; import org.junit.Test;
@ -152,7 +153,7 @@ public class TestBlockUnderConstruction {
*/ */
@Test @Test
public void testGetBlockLocations() throws IOException { public void testGetBlockLocations() throws IOException {
final NameNode namenode = cluster.getNameNode(); final NamenodeProtocols namenode = cluster.getNameNodeRpc();
final Path p = new Path(BASE_DIR, "file2.dat"); final Path p = new Path(BASE_DIR, "file2.dat");
final String src = p.toString(); final String src = p.toString();
final FSDataOutputStream out = TestFileCreation.createFile(hdfs, p, 3); final FSDataOutputStream out = TestFileCreation.createFile(hdfs, p, 3);

View File

@ -52,6 +52,7 @@ import org.apache.hadoop.hdfs.server.common.StorageInfo;
import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeDirType; import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeDirType;
import org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode.CheckpointStorage; import org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode.CheckpointStorage;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocol; import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocol;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.server.protocol.RemoteEditLog; import org.apache.hadoop.hdfs.server.protocol.RemoteEditLog;
import org.apache.hadoop.hdfs.server.protocol.RemoteEditLogManifest; import org.apache.hadoop.hdfs.server.protocol.RemoteEditLogManifest;
import org.apache.hadoop.hdfs.tools.DFSAdmin; import org.apache.hadoop.hdfs.tools.DFSAdmin;
@ -982,11 +983,12 @@ public class TestCheckpoint extends TestCase {
cluster = new MiniDFSCluster.Builder(conf).numDataNodes(numDatanodes) cluster = new MiniDFSCluster.Builder(conf).numDataNodes(numDatanodes)
.format(true).build(); .format(true).build();
NameNode nn = cluster.getNameNode(); NameNode nn = cluster.getNameNode();
NamenodeProtocols nnRpc = nn.getRpcServer();
SecondaryNameNode secondary = startSecondaryNameNode(conf); SecondaryNameNode secondary = startSecondaryNameNode(conf);
// prepare checkpoint image // prepare checkpoint image
secondary.doCheckpoint(); secondary.doCheckpoint();
CheckpointSignature sig = nn.rollEditLog(); CheckpointSignature sig = nnRpc.rollEditLog();
// manipulate the CheckpointSignature fields // manipulate the CheckpointSignature fields
sig.setBlockpoolID("somerandomebpid"); sig.setBlockpoolID("somerandomebpid");
sig.clusterID = "somerandomcid"; sig.clusterID = "somerandomcid";
@ -1073,8 +1075,10 @@ public class TestCheckpoint extends TestCase {
.nameNodePort(9928).build(); .nameNodePort(9928).build();
Configuration snConf1 = new HdfsConfiguration(cluster.getConfiguration(0)); Configuration snConf1 = new HdfsConfiguration(cluster.getConfiguration(0));
Configuration snConf2 = new HdfsConfiguration(cluster.getConfiguration(1)); Configuration snConf2 = new HdfsConfiguration(cluster.getConfiguration(1));
InetSocketAddress nn1RpcAddress = cluster.getNameNode(0).rpcAddress; InetSocketAddress nn1RpcAddress =
InetSocketAddress nn2RpcAddress = cluster.getNameNode(1).rpcAddress; cluster.getNameNode(0).getNameNodeAddress();
InetSocketAddress nn2RpcAddress =
cluster.getNameNode(1).getNameNodeAddress();
String nn1 = nn1RpcAddress.getHostName() + ":" + nn1RpcAddress.getPort(); String nn1 = nn1RpcAddress.getHostName() + ":" + nn1RpcAddress.getPort();
String nn2 = nn2RpcAddress.getHostName() + ":" + nn2RpcAddress.getPort(); String nn2 = nn2RpcAddress.getHostName() + ":" + nn2RpcAddress.getPort();
@ -1444,9 +1448,9 @@ public class TestCheckpoint extends TestCase {
cluster = new MiniDFSCluster.Builder(conf).numDataNodes(0) cluster = new MiniDFSCluster.Builder(conf).numDataNodes(0)
.format(true).build(); .format(true).build();
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
String fsName = NameNode.getHostPortString(nn.getHttpAddress()); String fsName = NameNode.getHostPortString(
cluster.getNameNode().getHttpAddress());
// Make a finalized log on the server side. // Make a finalized log on the server side.
nn.rollEditLog(); nn.rollEditLog();
@ -1515,8 +1519,8 @@ public class TestCheckpoint extends TestCase {
// Now primary NN experiences failure of a volume -- fake by // Now primary NN experiences failure of a volume -- fake by
// setting its current dir to a-x permissions // setting its current dir to a-x permissions
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
NNStorage storage = nn.getFSImage().getStorage(); NNStorage storage = cluster.getNameNode().getFSImage().getStorage();
StorageDirectory sd0 = storage.getStorageDir(0); StorageDirectory sd0 = storage.getStorageDir(0);
StorageDirectory sd1 = storage.getStorageDir(1); StorageDirectory sd1 = storage.getStorageDir(1);
@ -1590,8 +1594,8 @@ public class TestCheckpoint extends TestCase {
// Now primary NN experiences failure of its only name dir -- fake by // Now primary NN experiences failure of its only name dir -- fake by
// setting its current dir to a-x permissions // setting its current dir to a-x permissions
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
NNStorage storage = nn.getFSImage().getStorage(); NNStorage storage = cluster.getNameNode().getFSImage().getStorage();
StorageDirectory sd0 = storage.getStorageDir(0); StorageDirectory sd0 = storage.getStorageDir(0);
assertEquals(NameNodeDirType.IMAGE, sd0.getStorageDirType()); assertEquals(NameNodeDirType.IMAGE, sd0.getStorageDirType());
currentDir = sd0.getCurrentDir(); currentDir = sd0.getCurrentDir();
@ -1704,7 +1708,7 @@ public class TestCheckpoint extends TestCase {
secondary.doCheckpoint(); secondary.doCheckpoint();
// Now primary NN saves namespace 3 times // Now primary NN saves namespace 3 times
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
nn.setSafeMode(SafeModeAction.SAFEMODE_ENTER); nn.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
for (int i = 0; i < 3; i++) { for (int i = 0; i < 3; i++) {
nn.saveNamespace(); nn.saveNamespace();

View File

@ -102,7 +102,7 @@ public class TestDeadDatanode {
dn.shutdown(); dn.shutdown();
waitForDatanodeState(reg.getStorageID(), false, 20000); waitForDatanodeState(reg.getStorageID(), false, 20000);
DatanodeProtocol dnp = cluster.getNameNode(); DatanodeProtocol dnp = cluster.getNameNodeRpc();
Block[] blocks = new Block[] { new Block(0) }; Block[] blocks = new Block[] { new Block(0) };
String[] delHints = new String[] { "" }; String[] delHints = new String[] { "" };

View File

@ -47,6 +47,7 @@ import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock; import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.protocol.CorruptFileBlocks; import org.apache.hadoop.hdfs.protocol.CorruptFileBlocks;
import org.apache.hadoop.hdfs.protocol.LocatedBlocks; import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.tools.DFSck; import org.apache.hadoop.hdfs.tools.DFSck;
import org.apache.hadoop.io.IOUtils; import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.security.UserGroupInformation; import org.apache.hadoop.security.UserGroupInformation;
@ -481,7 +482,7 @@ public class TestFsck extends TestCase {
} }
// wait for the namenode to see the corruption // wait for the namenode to see the corruption
final NameNode namenode = cluster.getNameNode(); final NamenodeProtocols namenode = cluster.getNameNodeRpc();
CorruptFileBlocks corruptFileBlocks = namenode CorruptFileBlocks corruptFileBlocks = namenode
.listCorruptFileBlocks("/corruptData", null); .listCorruptFileBlocks("/corruptData", null);
int numCorrupt = corruptFileBlocks.getFiles().length; int numCorrupt = corruptFileBlocks.getFiles().length;

View File

@ -40,6 +40,7 @@ import org.apache.hadoop.hdfs.DistributedFileSystem;
import org.apache.hadoop.hdfs.MiniDFSCluster; import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.protocol.HdfsFileStatus; import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
import org.apache.hadoop.hdfs.protocol.LocatedBlocks; import org.apache.hadoop.hdfs.protocol.LocatedBlocks;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.tools.DFSAdmin; import org.apache.hadoop.hdfs.tools.DFSAdmin;
import org.apache.hadoop.security.UserGroupInformation; import org.apache.hadoop.security.UserGroupInformation;
import org.junit.After; import org.junit.After;
@ -52,7 +53,7 @@ public class TestHDFSConcat {
private static final short REPL_FACTOR = 2; private static final short REPL_FACTOR = 2;
private MiniDFSCluster cluster; private MiniDFSCluster cluster;
private NameNode nn; private NamenodeProtocols nn;
private DistributedFileSystem dfs; private DistributedFileSystem dfs;
private static long blockSize = 512; private static long blockSize = 512;
@ -72,7 +73,7 @@ public class TestHDFSConcat {
cluster.waitClusterUp(); cluster.waitClusterUp();
dfs = (DistributedFileSystem) cluster.getFileSystem(); dfs = (DistributedFileSystem) cluster.getFileSystem();
assertNotNull("Failed to get FileSystem", dfs); assertNotNull("Failed to get FileSystem", dfs);
nn = cluster.getNameNode(); nn = cluster.getNameNodeRpc();
assertNotNull("Failed to get NameNode", nn); assertNotNull("Failed to get NameNode", nn);
} }
@ -283,7 +284,7 @@ public class TestHDFSConcat {
Path filePath1 = new Path(name1); Path filePath1 = new Path(name1);
DFSTestUtil.createFile(dfs, filePath1, trgFileLen, REPL_FACTOR, 1); DFSTestUtil.createFile(dfs, filePath1, trgFileLen, REPL_FACTOR, 1);
HdfsFileStatus fStatus = cluster.getNameNode().getFileInfo(name1); HdfsFileStatus fStatus = nn.getFileInfo(name1);
long fileLen = fStatus.getLen(); long fileLen = fStatus.getLen();
assertEquals(fileLen, trgFileLen); assertEquals(fileLen, trgFileLen);
@ -293,11 +294,11 @@ public class TestHDFSConcat {
stm.readFully(0, byteFile1); stm.readFully(0, byteFile1);
stm.close(); stm.close();
LocatedBlocks lb1 = cluster.getNameNode().getBlockLocations(name1, 0, trgFileLen); LocatedBlocks lb1 = nn.getBlockLocations(name1, 0, trgFileLen);
Path filePath2 = new Path(name2); Path filePath2 = new Path(name2);
DFSTestUtil.createFile(dfs, filePath2, srcFileLen, REPL_FACTOR, 1); DFSTestUtil.createFile(dfs, filePath2, srcFileLen, REPL_FACTOR, 1);
fStatus = cluster.getNameNode().getFileInfo(name2); fStatus = nn.getFileInfo(name2);
fileLen = fStatus.getLen(); fileLen = fStatus.getLen();
assertEquals(srcFileLen, fileLen); assertEquals(srcFileLen, fileLen);
@ -307,7 +308,7 @@ public class TestHDFSConcat {
stm.readFully(0, byteFile2); stm.readFully(0, byteFile2);
stm.close(); stm.close();
LocatedBlocks lb2 = cluster.getNameNode().getBlockLocations(name2, 0, srcFileLen); LocatedBlocks lb2 = nn.getBlockLocations(name2, 0, srcFileLen);
System.out.println("trg len="+trgFileLen+"; src len="+srcFileLen); System.out.println("trg len="+trgFileLen+"; src len="+srcFileLen);
@ -316,7 +317,7 @@ public class TestHDFSConcat {
dfs.concat(filePath1, new Path [] {filePath2}); dfs.concat(filePath1, new Path [] {filePath2});
long totalLen = trgFileLen + srcFileLen; long totalLen = trgFileLen + srcFileLen;
fStatus = cluster.getNameNode().getFileInfo(name1); fStatus = nn.getFileInfo(name1);
fileLen = fStatus.getLen(); fileLen = fStatus.getLen();
// read the resulting file // read the resulting file
@ -325,7 +326,7 @@ public class TestHDFSConcat {
stm.readFully(0, byteFileConcat); stm.readFully(0, byteFileConcat);
stm.close(); stm.close();
LocatedBlocks lbConcat = cluster.getNameNode().getBlockLocations(name1, 0, fileLen); LocatedBlocks lbConcat = nn.getBlockLocations(name1, 0, fileLen);
//verifications //verifications
// 1. number of blocks // 1. number of blocks
@ -337,7 +338,7 @@ public class TestHDFSConcat {
assertEquals(fileLen, totalLen); assertEquals(fileLen, totalLen);
// 3. removal of the src file // 3. removal of the src file
fStatus = cluster.getNameNode().getFileInfo(name2); fStatus = nn.getFileInfo(name2);
assertNull("File "+name2+ "still exists", fStatus); // file shouldn't exist assertNull("File "+name2+ "still exists", fStatus); // file shouldn't exist
// 4. content // 4. content

View File

@ -147,7 +147,7 @@ public class TestListCorruptFileBlocks {
conf.setFloat(DFSConfigKeys.DFS_NAMENODE_REPL_QUEUE_THRESHOLD_PCT_KEY, conf.setFloat(DFSConfigKeys.DFS_NAMENODE_REPL_QUEUE_THRESHOLD_PCT_KEY,
0f); 0f);
cluster = new MiniDFSCluster.Builder(conf).waitSafeMode(false).build(); cluster = new MiniDFSCluster.Builder(conf).waitSafeMode(false).build();
cluster.getNameNode(). cluster.getNameNodeRpc().
setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
FileSystem fs = cluster.getFileSystem(); FileSystem fs = cluster.getFileSystem();
@ -244,7 +244,7 @@ public class TestListCorruptFileBlocks {
cluster.getNameNode().isInSafeMode()); cluster.getNameNode().isInSafeMode());
// now leave safe mode so that we can clean up // now leave safe mode so that we can clean up
cluster.getNameNode(). cluster.getNameNodeRpc().
setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE); setSafeMode(HdfsConstants.SafeModeAction.SAFEMODE_LEAVE);
util.cleanup(fs, "/srcdat10"); util.cleanup(fs, "/srcdat10");

View File

@ -148,8 +148,8 @@ public class TestNNStorageRetentionFunctional {
private static void doSaveNamespace(NameNode nn) throws IOException { private static void doSaveNamespace(NameNode nn) throws IOException {
LOG.info("Saving namespace..."); LOG.info("Saving namespace...");
nn.setSafeMode(SafeModeAction.SAFEMODE_ENTER); nn.getRpcServer().setSafeMode(SafeModeAction.SAFEMODE_ENTER);
nn.saveNamespace(); nn.getRpcServer().saveNamespace();
nn.setSafeMode(SafeModeAction.SAFEMODE_LEAVE); nn.getRpcServer().setSafeMode(SafeModeAction.SAFEMODE_LEAVE);
} }
} }

View File

@ -28,6 +28,7 @@ import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hdfs.DFSConfigKeys; import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.hdfs.HdfsConfiguration; import org.apache.hadoop.hdfs.HdfsConfiguration;
import org.apache.hadoop.hdfs.MiniDFSCluster; import org.apache.hadoop.hdfs.MiniDFSCluster;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.security.UserGroupInformation; import org.apache.hadoop.security.UserGroupInformation;
import org.junit.After; import org.junit.After;
import org.junit.Assert; import org.junit.Assert;
@ -54,7 +55,7 @@ public class TestNameNodeJspHelper {
@Test @Test
public void testDelegationToken() throws IOException, InterruptedException { public void testDelegationToken() throws IOException, InterruptedException {
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
HttpServletRequest request = mock(HttpServletRequest.class); HttpServletRequest request = mock(HttpServletRequest.class);
UserGroupInformation ugi = UserGroupInformation.createRemoteUser("auser"); UserGroupInformation ugi = UserGroupInformation.createRemoteUser("auser");
String tokenString = NamenodeJspHelper.getDelegationToken(nn, request, String tokenString = NamenodeJspHelper.getDelegationToken(nn, request,

View File

@ -108,7 +108,7 @@ public class TestParallelImageWrite extends TestCase {
files.cleanup(fs, dir); files.cleanup(fs, dir);
files.createFiles(fs, dir); files.createFiles(fs, dir);
fsn.setSafeMode(SafeModeAction.SAFEMODE_ENTER); fsn.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
cluster.getNameNode().saveNamespace(); cluster.getNameNodeRpc().saveNamespace();
final String checkAfterModify = checkImages(fsn, numNamenodeDirs); final String checkAfterModify = checkImages(fsn, numNamenodeDirs);
assertFalse("Modified namespace should change fsimage contents. " + assertFalse("Modified namespace should change fsimage contents. " +
"was: " + checkAfterRestart + " now: " + checkAfterModify, "was: " + checkAfterRestart + " now: " + checkAfterModify,

View File

@ -51,6 +51,7 @@ import org.apache.hadoop.hdfs.protocol.HdfsConstants.SafeModeAction;
import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory; import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory;
import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeDirType; import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeDirType;
import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeFile; import org.apache.hadoop.hdfs.server.namenode.NNStorage.NameNodeFile;
import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.util.MD5FileUtils; import org.apache.hadoop.hdfs.util.MD5FileUtils;
import org.apache.hadoop.io.MD5Hash; import org.apache.hadoop.io.MD5Hash;
import org.apache.hadoop.util.StringUtils; import org.apache.hadoop.util.StringUtils;
@ -379,9 +380,10 @@ public class TestStartup extends TestCase {
NameNode namenode = new NameNode(conf); NameNode namenode = new NameNode(conf);
namenode.getNamesystem().mkdirs("/test", namenode.getNamesystem().mkdirs("/test",
new PermissionStatus("hairong", null, FsPermission.getDefault()), true); new PermissionStatus("hairong", null, FsPermission.getDefault()), true);
assertTrue(namenode.getFileInfo("/test").isDir()); NamenodeProtocols nnRpc = namenode.getRpcServer();
namenode.setSafeMode(SafeModeAction.SAFEMODE_ENTER); assertTrue(nnRpc.getFileInfo("/test").isDir());
namenode.saveNamespace(); nnRpc.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
nnRpc.saveNamespace();
namenode.stop(); namenode.stop();
namenode.join(); namenode.join();
@ -408,9 +410,10 @@ public class TestStartup extends TestCase {
private void checkNameSpace(Configuration conf) throws IOException { private void checkNameSpace(Configuration conf) throws IOException {
NameNode namenode = new NameNode(conf); NameNode namenode = new NameNode(conf);
assertTrue(namenode.getFileInfo("/test").isDir()); NamenodeProtocols nnRpc = namenode.getRpcServer();
namenode.setSafeMode(SafeModeAction.SAFEMODE_ENTER); assertTrue(nnRpc.getFileInfo("/test").isDir());
namenode.saveNamespace(); nnRpc.setSafeMode(SafeModeAction.SAFEMODE_ENTER);
nnRpc.saveNamespace();
namenode.stop(); namenode.stop();
namenode.join(); namenode.join();
} }
@ -515,7 +518,7 @@ public class TestStartup extends TestCase {
cluster.waitActive(); cluster.waitActive();
cluster.restartNameNode(); cluster.restartNameNode();
NameNode nn = cluster.getNameNode(); NamenodeProtocols nn = cluster.getNameNodeRpc();
assertNotNull(nn); assertNotNull(nn);
Assert.assertTrue(cluster.isDataNodeUp()); Assert.assertTrue(cluster.isDataNodeUp());

View File

@ -339,7 +339,7 @@ public class TestStorageRestore {
// Simulate a 2NN beginning a checkpoint, but not finishing. This will // Simulate a 2NN beginning a checkpoint, but not finishing. This will
// cause name1 to be restored. // cause name1 to be restored.
cluster.getNameNode().rollEditLog(); cluster.getNameNodeRpc().rollEditLog();
printStorages(fsImage); printStorages(fsImage);

View File

@ -74,9 +74,9 @@ public class TestNNMetricFilesInGetListingOps extends TestCase {
createFile("/tmp1/t2", 3200, (short)3); createFile("/tmp1/t2", 3200, (short)3);
createFile("/tmp2/t1", 3200, (short)3); createFile("/tmp2/t1", 3200, (short)3);
createFile("/tmp2/t2", 3200, (short)3); createFile("/tmp2/t2", 3200, (short)3);
cluster.getNameNode().getListing("/tmp1", HdfsFileStatus.EMPTY_NAME, false); cluster.getNameNodeRpc().getListing("/tmp1", HdfsFileStatus.EMPTY_NAME, false);
assertCounter("FilesInGetListingOps", 2L, getMetrics(NN_METRICS)); assertCounter("FilesInGetListingOps", 2L, getMetrics(NN_METRICS));
cluster.getNameNode().getListing("/tmp2", HdfsFileStatus.EMPTY_NAME, false); cluster.getNameNodeRpc().getListing("/tmp2", HdfsFileStatus.EMPTY_NAME, false);
assertCounter("FilesInGetListingOps", 4L, getMetrics(NN_METRICS)); assertCounter("FilesInGetListingOps", 4L, getMetrics(NN_METRICS));
} }
} }

View File

@ -124,8 +124,8 @@ public class TestOfflineImageViewer extends TestCase {
} }
// Write results to the fsimage file // Write results to the fsimage file
cluster.getNameNode().setSafeMode(SafeModeAction.SAFEMODE_ENTER); cluster.getNameNodeRpc().setSafeMode(SafeModeAction.SAFEMODE_ENTER);
cluster.getNameNode().saveNamespace(); cluster.getNameNodeRpc().saveNamespace();
// Determine location of fsimage file // Determine location of fsimage file
orig = FSImageTestUtil.findLatestImageFile( orig = FSImageTestUtil.findLatestImageFile(