HDFS-2147. Move cluster network topology to block management and fix some javac warnings.
git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1148112 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
parent
2a49e1a3a4
commit
c3f6575ca4
@ -569,6 +569,9 @@ Trunk (unreleased changes)
|
||||
HDFS-2157. Improve header comment in o.a.h.hdfs.server.namenode.NameNode.
|
||||
(atm via eli)
|
||||
|
||||
HDFS-2147. Move cluster network topology to block management and fix some
|
||||
javac warnings. (szetszwo)
|
||||
|
||||
OPTIMIZATIONS
|
||||
|
||||
HDFS-1458. Improve checkpoint performance by avoiding unnecessary image
|
||||
|
@ -22,6 +22,7 @@
|
||||
import java.util.ArrayList;
|
||||
import java.util.Collection;
|
||||
import java.util.Collections;
|
||||
import java.util.HashMap;
|
||||
import java.util.HashSet;
|
||||
import java.util.Iterator;
|
||||
import java.util.LinkedList;
|
||||
@ -31,6 +32,8 @@
|
||||
import java.util.TreeMap;
|
||||
import java.util.TreeSet;
|
||||
|
||||
import org.apache.commons.logging.Log;
|
||||
import org.apache.commons.logging.LogFactory;
|
||||
import org.apache.hadoop.classification.InterfaceAudience;
|
||||
import org.apache.hadoop.conf.Configuration;
|
||||
import org.apache.hadoop.hdfs.DFSConfigKeys;
|
||||
@ -47,6 +50,7 @@
|
||||
import org.apache.hadoop.hdfs.server.namenode.INodeFile;
|
||||
import org.apache.hadoop.hdfs.server.namenode.INodeFileUnderConstruction;
|
||||
import org.apache.hadoop.hdfs.server.namenode.NameNode;
|
||||
import org.apache.hadoop.net.Node;
|
||||
|
||||
/**
|
||||
* Keeps information related to the blocks stored in the Hadoop cluster.
|
||||
@ -55,8 +59,9 @@
|
||||
*/
|
||||
@InterfaceAudience.Private
|
||||
public class BlockManager {
|
||||
// Default initial capacity and load factor of map
|
||||
public static final int DEFAULT_INITIAL_MAP_CAPACITY = 16;
|
||||
static final Log LOG = LogFactory.getLog(BlockManager.class);
|
||||
|
||||
/** Default load factor of map */
|
||||
public static final float DEFAULT_MAP_LOAD_FACTOR = 0.75f;
|
||||
|
||||
private final FSNamesystem namesystem;
|
||||
@ -104,7 +109,7 @@ public long getExcessBlocksCount() {
|
||||
//
|
||||
// Store blocks-->datanodedescriptor(s) map of corrupt replicas
|
||||
//
|
||||
CorruptReplicasMap corruptReplicas = new CorruptReplicasMap();
|
||||
private final CorruptReplicasMap corruptReplicas = new CorruptReplicasMap();
|
||||
|
||||
//
|
||||
// Keeps a Collection for every named machine containing
|
||||
@ -112,7 +117,7 @@ public long getExcessBlocksCount() {
|
||||
// on the machine in question.
|
||||
// Mapping: StorageID -> ArrayList<Block>
|
||||
//
|
||||
Map<String, Collection<Block>> recentInvalidateSets =
|
||||
private final Map<String, Collection<Block>> recentInvalidateSets =
|
||||
new TreeMap<String, Collection<Block>>();
|
||||
|
||||
//
|
||||
@ -128,22 +133,22 @@ public long getExcessBlocksCount() {
|
||||
// Store set of Blocks that need to be replicated 1 or more times.
|
||||
// We also store pending replication-orders.
|
||||
//
|
||||
public UnderReplicatedBlocks neededReplications = new UnderReplicatedBlocks();
|
||||
private PendingReplicationBlocks pendingReplications;
|
||||
public final UnderReplicatedBlocks neededReplications = new UnderReplicatedBlocks();
|
||||
private final PendingReplicationBlocks pendingReplications;
|
||||
|
||||
// The maximum number of replicas allowed for a block
|
||||
public int maxReplication;
|
||||
public final int maxReplication;
|
||||
// How many outgoing replication streams a given node should have at one time
|
||||
public int maxReplicationStreams;
|
||||
// Minimum copies needed or else write is disallowed
|
||||
public int minReplication;
|
||||
public final int minReplication;
|
||||
// Default number of replicas
|
||||
public int defaultReplication;
|
||||
public final int defaultReplication;
|
||||
// How many entries are returned by getCorruptInodes()
|
||||
int maxCorruptFilesReturned;
|
||||
final int maxCorruptFilesReturned;
|
||||
|
||||
// variable to enable check for enough racks
|
||||
boolean shouldCheckForEnoughRacks = true;
|
||||
final boolean shouldCheckForEnoughRacks;
|
||||
|
||||
/**
|
||||
* Last block index used for replication work.
|
||||
@ -152,28 +157,18 @@ public long getExcessBlocksCount() {
|
||||
Random r = new Random();
|
||||
|
||||
// for block replicas placement
|
||||
public BlockPlacementPolicy replicator;
|
||||
public final BlockPlacementPolicy replicator;
|
||||
|
||||
public BlockManager(FSNamesystem fsn, Configuration conf) throws IOException {
|
||||
this(fsn, conf, DEFAULT_INITIAL_MAP_CAPACITY);
|
||||
}
|
||||
|
||||
BlockManager(FSNamesystem fsn, Configuration conf, int capacity)
|
||||
throws IOException {
|
||||
namesystem = fsn;
|
||||
datanodeManager = new DatanodeManager(fsn);
|
||||
|
||||
blocksMap = new BlocksMap(DEFAULT_MAP_LOAD_FACTOR);
|
||||
replicator = BlockPlacementPolicy.getInstance(
|
||||
conf, namesystem, datanodeManager.getNetworkTopology());
|
||||
pendingReplications = new PendingReplicationBlocks(conf.getInt(
|
||||
DFSConfigKeys.DFS_NAMENODE_REPLICATION_PENDING_TIMEOUT_SEC_KEY,
|
||||
DFSConfigKeys.DFS_NAMENODE_REPLICATION_PENDING_TIMEOUT_SEC_DEFAULT) * 1000L);
|
||||
setConfigurationParameters(conf);
|
||||
blocksMap = new BlocksMap(capacity, DEFAULT_MAP_LOAD_FACTOR);
|
||||
datanodeManager = new DatanodeManager(fsn);
|
||||
}
|
||||
|
||||
void setConfigurationParameters(Configuration conf) throws IOException {
|
||||
this.replicator = BlockPlacementPolicy.getInstance(
|
||||
conf,
|
||||
namesystem,
|
||||
namesystem.clusterMap);
|
||||
|
||||
this.maxCorruptFilesReturned = conf.getInt(
|
||||
DFSConfigKeys.DFS_DEFAULT_MAX_CORRUPT_FILES_RETURNED_KEY,
|
||||
@ -541,6 +536,22 @@ public void verifyReplication(String src,
|
||||
minReplication);
|
||||
}
|
||||
|
||||
/** Remove a datanode. */
|
||||
public void removeDatanode(final DatanodeDescriptor node) {
|
||||
final Iterator<? extends Block> it = node.getBlockIterator();
|
||||
while(it.hasNext()) {
|
||||
removeStoredBlock(it.next(), node);
|
||||
}
|
||||
|
||||
node.resetBlocks();
|
||||
removeFromInvalidates(node.getStorageID());
|
||||
datanodeManager.getNetworkTopology().remove(node);
|
||||
|
||||
if (LOG.isDebugEnabled()) {
|
||||
LOG.debug("remove datanode " + node.getName());
|
||||
}
|
||||
}
|
||||
|
||||
void removeFromInvalidates(String storageID, Block block) {
|
||||
Collection<Block> v = recentInvalidateSets.get(storageID);
|
||||
if (v != null && v.remove(block)) {
|
||||
@ -1001,6 +1012,29 @@ private boolean computeReplicationWorkForBlock(Block block, int priority) {
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Choose target datanodes according to the replication policy.
|
||||
* @throws IOException if the number of targets < minimum replication.
|
||||
* @see BlockPlacementPolicy#chooseTarget(String, int, DatanodeDescriptor, HashMap, long)
|
||||
*/
|
||||
public DatanodeDescriptor[] chooseTarget(final String src,
|
||||
final int numOfReplicas, final DatanodeDescriptor client,
|
||||
final HashMap<Node, Node> excludedNodes,
|
||||
final long blocksize) throws IOException {
|
||||
// choose targets for the new block to be allocated.
|
||||
final DatanodeDescriptor targets[] = replicator.chooseTarget(
|
||||
src, numOfReplicas, client, excludedNodes, blocksize);
|
||||
if (targets.length < minReplication) {
|
||||
throw new IOException("File " + src + " could only be replicated to " +
|
||||
targets.length + " nodes, instead of " +
|
||||
minReplication + ". There are "
|
||||
+ getDatanodeManager().getNetworkTopology().getNumOfLeaves()
|
||||
+ " datanode(s) running but "+excludedNodes.size() +
|
||||
" node(s) are excluded in this operation.");
|
||||
}
|
||||
return targets;
|
||||
}
|
||||
|
||||
/**
|
||||
* Parse the data-nodes the block belongs to and choose one,
|
||||
* which will be the replication source.
|
||||
|
@ -57,7 +57,7 @@ public void remove() {
|
||||
|
||||
private GSet<Block, BlockInfo> blocks;
|
||||
|
||||
BlocksMap(int initialCapacity, float loadFactor) {
|
||||
BlocksMap(final float loadFactor) {
|
||||
this.capacity = computeCapacity();
|
||||
this.blocks = new LightWeightGSet<Block, BlockInfo>(capacity);
|
||||
}
|
||||
|
@ -18,6 +18,8 @@
|
||||
package org.apache.hadoop.hdfs.server.blockmanagement;
|
||||
|
||||
import java.io.IOException;
|
||||
import java.util.Arrays;
|
||||
import java.util.List;
|
||||
|
||||
import org.apache.commons.logging.Log;
|
||||
import org.apache.commons.logging.LogFactory;
|
||||
@ -25,8 +27,11 @@
|
||||
import org.apache.hadoop.classification.InterfaceStability;
|
||||
import org.apache.hadoop.conf.Configuration;
|
||||
import org.apache.hadoop.hdfs.DFSConfigKeys;
|
||||
import org.apache.hadoop.hdfs.DFSUtil;
|
||||
import org.apache.hadoop.hdfs.protocol.DatanodeID;
|
||||
import org.apache.hadoop.hdfs.protocol.LocatedBlock;
|
||||
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
|
||||
import org.apache.hadoop.net.NetworkTopology;
|
||||
import org.apache.hadoop.util.Daemon;
|
||||
|
||||
/**
|
||||
@ -39,6 +44,10 @@ public class DatanodeManager {
|
||||
|
||||
final FSNamesystem namesystem;
|
||||
|
||||
/** Cluster network topology */
|
||||
private final NetworkTopology networktopology = new NetworkTopology();
|
||||
|
||||
/** Host names to datanode descriptors mapping. */
|
||||
private final Host2NodesMap host2DatanodeMap = new Host2NodesMap();
|
||||
|
||||
DatanodeManager(final FSNamesystem namesystem) {
|
||||
@ -60,6 +69,24 @@ void close() {
|
||||
if (decommissionthread != null) decommissionthread.interrupt();
|
||||
}
|
||||
|
||||
/** @return the network topology. */
|
||||
public NetworkTopology getNetworkTopology() {
|
||||
return networktopology;
|
||||
}
|
||||
|
||||
/** Sort the located blocks by the distance to the target host. */
|
||||
public void sortLocatedBlocks(final String targethost,
|
||||
final List<LocatedBlock> locatedblocks) {
|
||||
//sort the blocks
|
||||
final DatanodeDescriptor client = getDatanodeByHost(targethost);
|
||||
for (LocatedBlock b : locatedblocks) {
|
||||
networktopology.pseudoSortByDistance(client, b.getLocations());
|
||||
|
||||
// Move decommissioned datanodes to the bottom
|
||||
Arrays.sort(b.getLocations(), DFSUtil.DECOM_COMPARATOR);
|
||||
}
|
||||
}
|
||||
|
||||
/** @return the datanode descriptor for the host. */
|
||||
public DatanodeDescriptor getDatanodeByHost(final String host) {
|
||||
return host2DatanodeMap.getDatanodeByHost(host);
|
||||
@ -74,10 +101,12 @@ public void addDatanode(final DatanodeDescriptor node) {
|
||||
host2DatanodeMap.remove(
|
||||
namesystem.datanodeMap.put(node.getStorageID(), node));
|
||||
}
|
||||
|
||||
host2DatanodeMap.add(node);
|
||||
networktopology.add(node);
|
||||
|
||||
if (LOG.isDebugEnabled()) {
|
||||
LOG.debug(getClass().getSimpleName() + ".unprotectedAddDatanode: "
|
||||
LOG.debug(getClass().getSimpleName() + ".addDatanode: "
|
||||
+ "node " + node.getName() + " is added to datanodeMap.");
|
||||
}
|
||||
}
|
||||
|
@ -138,7 +138,6 @@
|
||||
import org.apache.hadoop.net.DNSToSwitchMapping;
|
||||
import org.apache.hadoop.net.NetworkTopology;
|
||||
import org.apache.hadoop.net.Node;
|
||||
import org.apache.hadoop.net.NodeBase;
|
||||
import org.apache.hadoop.net.ScriptBasedMapping;
|
||||
import org.apache.hadoop.security.AccessControlException;
|
||||
import org.apache.hadoop.security.UserGroupInformation;
|
||||
@ -240,7 +239,7 @@ private static final void logAuditEvent(UserGroupInformation ugi,
|
||||
// Stores the correct file name hierarchy
|
||||
//
|
||||
public FSDirectory dir;
|
||||
public BlockManager blockManager;
|
||||
BlockManager blockManager;
|
||||
|
||||
// Block pool ID used by this namenode
|
||||
String blockPoolId;
|
||||
@ -318,8 +317,6 @@ private static final void logAuditEvent(UserGroupInformation ugi,
|
||||
|
||||
private volatile SafeModeInfo safeMode; // safe mode information
|
||||
|
||||
/** datanode network toplogy */
|
||||
public NetworkTopology clusterMap = new NetworkTopology();
|
||||
private DNSToSwitchMapping dnsToSwitchMapping;
|
||||
|
||||
private HostsFileReader hostsReader;
|
||||
@ -876,15 +873,8 @@ LocatedBlocks getBlockLocations(String clientMachine, String src,
|
||||
FileNotFoundException, UnresolvedLinkException, IOException {
|
||||
LocatedBlocks blocks = getBlockLocations(src, offset, length, true, true);
|
||||
if (blocks != null) {
|
||||
//sort the blocks
|
||||
final DatanodeDescriptor client =
|
||||
blockManager.getDatanodeManager().getDatanodeByHost(clientMachine);
|
||||
for (LocatedBlock b : blocks.getLocatedBlocks()) {
|
||||
clusterMap.pseudoSortByDistance(client, b.getLocations());
|
||||
|
||||
// Move decommissioned datanodes to the bottom
|
||||
Arrays.sort(b.getLocations(), DFSUtil.DECOM_COMPARATOR);
|
||||
}
|
||||
blockManager.getDatanodeManager().sortLocatedBlocks(
|
||||
clientMachine, blocks.getLocatedBlocks());
|
||||
}
|
||||
return blocks;
|
||||
}
|
||||
@ -1774,16 +1764,8 @@ public LocatedBlock getAdditionalBlock(String src,
|
||||
}
|
||||
|
||||
// choose targets for the new block to be allocated.
|
||||
DatanodeDescriptor targets[] = blockManager.replicator.chooseTarget(
|
||||
final DatanodeDescriptor targets[] = blockManager.chooseTarget(
|
||||
src, replication, clientNode, excludedNodes, blockSize);
|
||||
if (targets.length < blockManager.minReplication) {
|
||||
throw new IOException("File " + src + " could only be replicated to " +
|
||||
targets.length + " nodes, instead of " +
|
||||
blockManager.minReplication + ". There are "
|
||||
+clusterMap.getNumOfLeaves()+" datanode(s) running"
|
||||
+" but "+excludedNodes.size() +
|
||||
" node(s) are excluded in this operation.");
|
||||
}
|
||||
|
||||
// Allocate a new block and record it in the INode.
|
||||
writeLock();
|
||||
@ -2881,14 +2863,14 @@ nodes with its data cleared (or user can just remove the StorageID
|
||||
nodeReg.getStorageID());
|
||||
}
|
||||
// update cluster map
|
||||
clusterMap.remove(nodeS);
|
||||
blockManager.getDatanodeManager().getNetworkTopology().remove(nodeS);
|
||||
nodeS.updateRegInfo(nodeReg);
|
||||
nodeS.setHostName(hostName);
|
||||
nodeS.setDisallowed(false); // Node is in the include list
|
||||
|
||||
// resolve network location
|
||||
resolveNetworkLocation(nodeS);
|
||||
clusterMap.add(nodeS);
|
||||
blockManager.getDatanodeManager().getNetworkTopology().add(nodeS);
|
||||
|
||||
// also treat the registration message as a heartbeat
|
||||
synchronized(heartbeats) {
|
||||
@ -2919,7 +2901,6 @@ nodes with its data cleared (or user can just remove the StorageID
|
||||
= new DatanodeDescriptor(nodeReg, NetworkTopology.DEFAULT_RACK, hostName);
|
||||
resolveNetworkLocation(nodeDescr);
|
||||
blockManager.getDatanodeManager().addDatanode(nodeDescr);
|
||||
clusterMap.add(nodeDescr);
|
||||
checkDecommissioning(nodeDescr, dnAddress);
|
||||
|
||||
// also treat the registration message as a heartbeat
|
||||
@ -3336,27 +3317,11 @@ private void removeDatanode(DatanodeDescriptor nodeInfo) {
|
||||
}
|
||||
}
|
||||
|
||||
Iterator<? extends Block> it = nodeInfo.getBlockIterator();
|
||||
while(it.hasNext()) {
|
||||
blockManager.removeStoredBlock(it.next(), nodeInfo);
|
||||
}
|
||||
unprotectedRemoveDatanode(nodeInfo);
|
||||
clusterMap.remove(nodeInfo);
|
||||
blockManager.removeDatanode(nodeInfo);
|
||||
|
||||
checkSafeMode();
|
||||
}
|
||||
|
||||
void unprotectedRemoveDatanode(DatanodeDescriptor nodeDescr) {
|
||||
assert hasWriteLock();
|
||||
nodeDescr.resetBlocks();
|
||||
blockManager.removeFromInvalidates(nodeDescr.getStorageID());
|
||||
if(NameNode.stateChangeLog.isDebugEnabled()) {
|
||||
NameNode.stateChangeLog.debug(
|
||||
"BLOCK* NameSystem.unprotectedRemoveDatanode: "
|
||||
+ nodeDescr.getName() + " is out of service now.");
|
||||
}
|
||||
}
|
||||
|
||||
FSImage getFSImage() {
|
||||
return dir.fsImage;
|
||||
}
|
||||
@ -4104,14 +4069,6 @@ public DatanodeDescriptor getDatanode(DatanodeID nodeID) throws IOException {
|
||||
return node;
|
||||
}
|
||||
|
||||
/** Choose a random datanode
|
||||
*
|
||||
* @return a randomly chosen datanode
|
||||
*/
|
||||
DatanodeDescriptor getRandomDatanode() {
|
||||
return (DatanodeDescriptor)clusterMap.chooseRandom(NodeBase.ROOT);
|
||||
}
|
||||
|
||||
/**
|
||||
* SafeModeInfo contains information related to the safe mode.
|
||||
* <p>
|
||||
@ -4278,9 +4235,10 @@ synchronized void leave(boolean checkForUpgrades) {
|
||||
}
|
||||
reached = -1;
|
||||
safeMode = null;
|
||||
final NetworkTopology nt = blockManager.getDatanodeManager().getNetworkTopology();
|
||||
NameNode.stateChangeLog.info("STATE* Network topology has "
|
||||
+clusterMap.getNumOfRacks()+" racks and "
|
||||
+clusterMap.getNumOfLeaves()+ " datanodes");
|
||||
+ nt.getNumOfRacks() + " racks and "
|
||||
+ nt.getNumOfLeaves() + " datanodes");
|
||||
NameNode.stateChangeLog.info("STATE* UnderReplicatedBlocks has "
|
||||
+blockManager.neededReplications.size()+" blocks");
|
||||
}
|
||||
@ -5866,6 +5824,11 @@ public String getBlockPoolId() {
|
||||
return blockPoolId;
|
||||
}
|
||||
|
||||
/** @return the block manager. */
|
||||
public BlockManager getBlockManager() {
|
||||
return blockManager;
|
||||
}
|
||||
|
||||
/**
|
||||
* Remove an already decommissioned data node who is neither in include nor
|
||||
* exclude hosts lists from the the list of live or dead nodes. This is used
|
||||
|
@ -32,14 +32,14 @@
|
||||
import org.apache.hadoop.conf.Configuration;
|
||||
import org.apache.hadoop.fs.MD5MD5CRC32FileChecksum;
|
||||
import org.apache.hadoop.hdfs.DFSClient;
|
||||
import org.apache.hadoop.hdfs.DFSConfigKeys;
|
||||
import org.apache.hadoop.hdfs.HdfsConfiguration;
|
||||
import org.apache.hadoop.hdfs.protocol.ClientProtocol;
|
||||
import org.apache.hadoop.hdfs.protocol.DatanodeID;
|
||||
import org.apache.hadoop.hdfs.server.common.HdfsConstants;
|
||||
import org.apache.hadoop.hdfs.server.common.JspHelper;
|
||||
import org.apache.hadoop.hdfs.server.datanode.DataNode;
|
||||
import org.apache.hadoop.hdfs.server.datanode.DatanodeJspHelper;
|
||||
import org.apache.hadoop.hdfs.DFSConfigKeys;
|
||||
import org.apache.hadoop.hdfs.HdfsConfiguration;
|
||||
import org.apache.hadoop.net.NetUtils;
|
||||
import org.apache.hadoop.security.UserGroupInformation;
|
||||
import org.znerd.xmlenc.XMLOutputter;
|
||||
@ -61,7 +61,7 @@ public void doGet(HttpServletRequest request, HttpServletResponse response
|
||||
(Configuration) context.getAttribute(JspHelper.CURRENT_CONF);
|
||||
final UserGroupInformation ugi = getUGI(request, conf);
|
||||
final NameNode namenode = (NameNode)context.getAttribute("name.node");
|
||||
final DatanodeID datanode = namenode.getNamesystem().getRandomDatanode();
|
||||
final DatanodeID datanode = NamenodeJspHelper.getRandomDatanode(namenode);
|
||||
try {
|
||||
final URI uri = createRedirectUri("/getFileChecksum", ugi, datanode,
|
||||
request, namenode);
|
||||
|
@ -86,7 +86,7 @@ private DatanodeID pickSrcDatanode(LocatedBlocks blks, HdfsFileStatus i)
|
||||
if (i.getLen() == 0 || blks.getLocatedBlocks().size() <= 0) {
|
||||
// pick a random datanode
|
||||
NameNode nn = (NameNode)getServletContext().getAttribute("name.node");
|
||||
return nn.getNamesystem().getRandomDatanode();
|
||||
return NamenodeJspHelper.getRandomDatanode(nn);
|
||||
}
|
||||
return JspHelper.bestNode(blks);
|
||||
}
|
||||
|
@ -66,7 +66,8 @@ public Object run() throws Exception {
|
||||
namesystem.getNumberOfDatanodes(DatanodeReportType.LIVE);
|
||||
final short minReplication = namesystem.getMinReplication();
|
||||
|
||||
new NamenodeFsck(conf, nn, nn.getNetworkTopology(), pmap, out,
|
||||
new NamenodeFsck(conf, nn,
|
||||
NamenodeJspHelper.getNetworkTopology(nn), pmap, out,
|
||||
totalDatanodes, minReplication, remoteAddress).fsck();
|
||||
|
||||
return null;
|
||||
|
@ -89,7 +89,6 @@
|
||||
import org.apache.hadoop.ipc.Server;
|
||||
import org.apache.hadoop.metrics2.lib.DefaultMetricsSystem;
|
||||
import org.apache.hadoop.net.NetUtils;
|
||||
import org.apache.hadoop.net.NetworkTopology;
|
||||
import org.apache.hadoop.net.Node;
|
||||
import org.apache.hadoop.security.AccessControlException;
|
||||
import org.apache.hadoop.security.Groups;
|
||||
@ -1424,10 +1423,6 @@ public InetSocketAddress getHttpAddress() {
|
||||
return httpAddress;
|
||||
}
|
||||
|
||||
NetworkTopology getNetworkTopology() {
|
||||
return this.namesystem.clusterMap;
|
||||
}
|
||||
|
||||
/**
|
||||
* Verify that configured directories exist, then
|
||||
* Interactively confirm that formatting is desired
|
||||
|
@ -48,6 +48,8 @@
|
||||
import org.apache.hadoop.hdfs.server.common.Storage.StorageDirectory;
|
||||
import org.apache.hadoop.hdfs.server.common.UpgradeStatusReport;
|
||||
import org.apache.hadoop.io.Text;
|
||||
import org.apache.hadoop.net.NetworkTopology;
|
||||
import org.apache.hadoop.net.NodeBase;
|
||||
import org.apache.hadoop.security.UserGroupInformation;
|
||||
import org.apache.hadoop.security.token.Token;
|
||||
import org.apache.hadoop.util.ServletUtil;
|
||||
@ -368,13 +370,25 @@ public Token<DelegationTokenIdentifier> run() throws IOException {
|
||||
return token == null ? null : token.encodeToUrlString();
|
||||
}
|
||||
|
||||
/** @return the network topology. */
|
||||
static NetworkTopology getNetworkTopology(final NameNode namenode) {
|
||||
return namenode.getNamesystem().getBlockManager().getDatanodeManager(
|
||||
).getNetworkTopology();
|
||||
}
|
||||
|
||||
/** @return a randomly chosen datanode. */
|
||||
static DatanodeDescriptor getRandomDatanode(final NameNode namenode) {
|
||||
return (DatanodeDescriptor)getNetworkTopology(namenode).chooseRandom(
|
||||
NodeBase.ROOT);
|
||||
}
|
||||
|
||||
static void redirectToRandomDataNode(ServletContext context,
|
||||
HttpServletRequest request, HttpServletResponse resp) throws IOException,
|
||||
InterruptedException {
|
||||
final NameNode nn = (NameNode) context.getAttribute("name.node");
|
||||
final Configuration conf = (Configuration) context
|
||||
.getAttribute(JspHelper.CURRENT_CONF);
|
||||
final DatanodeID datanode = nn.getNamesystem().getRandomDatanode();
|
||||
final DatanodeID datanode = getRandomDatanode(nn);
|
||||
UserGroupInformation ugi = JspHelper.getUGI(context, request, conf);
|
||||
String tokenString = getDelegationToken(nn, request, conf, ugi);
|
||||
// if the user is defined, get a delegation token and stringify it
|
||||
|
@ -32,7 +32,6 @@
|
||||
import org.apache.hadoop.hdfs.HdfsConfiguration;
|
||||
import org.apache.hadoop.hdfs.protocol.DatanodeID;
|
||||
import org.apache.hadoop.hdfs.protocol.FSConstants;
|
||||
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
|
||||
import org.apache.hadoop.hdfs.server.namenode.NameNode;
|
||||
import org.apache.hadoop.net.NetworkTopology;
|
||||
import org.apache.hadoop.net.Node;
|
||||
@ -68,9 +67,9 @@ public class TestReplicationPolicy extends TestCase {
|
||||
e.printStackTrace();
|
||||
throw (RuntimeException)new RuntimeException().initCause(e);
|
||||
}
|
||||
FSNamesystem fsNamesystem = namenode.getNamesystem();
|
||||
replicator = fsNamesystem.blockManager.replicator;
|
||||
cluster = fsNamesystem.clusterMap;
|
||||
final BlockManager bm = namenode.getNamesystem().getBlockManager();
|
||||
replicator = bm.replicator;
|
||||
cluster = bm.getDatanodeManager().getNetworkTopology();
|
||||
// construct network topology
|
||||
for(int i=0; i<NUM_OF_DATANODES; i++) {
|
||||
cluster.add(dataNodes[i]);
|
||||
|
@ -27,7 +27,6 @@
|
||||
import org.apache.hadoop.hdfs.HdfsConfiguration;
|
||||
import org.apache.hadoop.hdfs.MiniDFSCluster;
|
||||
import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
|
||||
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
|
||||
|
||||
public class TestUnderReplicatedBlocks extends TestCase {
|
||||
public void testSetrepIncWithUnderReplicatedBlocks() throws Exception {
|
||||
@ -44,11 +43,11 @@ public void testSetrepIncWithUnderReplicatedBlocks() throws Exception {
|
||||
|
||||
// remove one replica from the blocksMap so block becomes under-replicated
|
||||
// but the block does not get put into the under-replicated blocks queue
|
||||
final FSNamesystem namesystem = cluster.getNamesystem();
|
||||
final BlockManager bm = cluster.getNamesystem().getBlockManager();
|
||||
ExtendedBlock b = DFSTestUtil.getFirstBlock(fs, FILE_PATH);
|
||||
DatanodeDescriptor dn = namesystem.blockManager.blocksMap.nodeIterator(b.getLocalBlock()).next();
|
||||
namesystem.blockManager.addToInvalidates(b.getLocalBlock(), dn);
|
||||
namesystem.blockManager.blocksMap.removeNode(b.getLocalBlock(), dn);
|
||||
DatanodeDescriptor dn = bm.blocksMap.nodeIterator(b.getLocalBlock()).next();
|
||||
bm.addToInvalidates(b.getLocalBlock(), dn);
|
||||
bm.blocksMap.removeNode(b.getLocalBlock(), dn);
|
||||
|
||||
// increment this file's replication factor
|
||||
FsShell shell = new FsShell(conf);
|
||||
|
Loading…
Reference in New Issue
Block a user