Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HDFS-7993. Provide each Replica details in fsck (Contributed by J.Andreina) 

(cherry picked from commit 8ddbb8dd43)
This commit is contained in:
Vinayakumar B 2015-04-21 15:24:49 +05:30
parent 27fc4fb998
commit e7bb0fc922
5 changed files with 132 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
View File

@ -208,6 +208,8 @@ Release 2.8.0 - UNRELEASED
HDFS-8173. NPE thrown at DataNode shutdown when HTTP server was not able to
create (surendra singh lilhore via vinayakumarb)
HDFS-7993. Provide each Replica details in fsck (J.Andreina via vinayakumarb)
Release 2.7.1 - UNRELEASED
INCOMPATIBLE CHANGES

6
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/DatanodeStorageInfo.java
View File

@ -155,7 +155,7 @@ public class DatanodeStorageInfo {
this.blockReportCount = blockReportCount;
}
boolean areBlockContentsStale() {
public boolean areBlockContentsStale() {
return blockContentsStale;
}
@ -205,11 +205,11 @@ public class DatanodeStorageInfo {
return getState() == State.FAILED && numBlocks != 0;
}
String getStorageID() {
public String getStorageID() {
return storageID;
}
StorageType getStorageType() {
public StorageType getStorageType() {
return storageType;
}

61
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/NamenodeFsck.java
View File

@ -52,6 +52,7 @@ import org.apache.hadoop.hdfs.net.TcpPeerServer;
import org.apache.hadoop.hdfs.protocol.Block;
import org.apache.hadoop.hdfs.protocol.DatanodeID;
import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
import org.apache.hadoop.hdfs.protocol.DatanodeInfoWithStorage;
import org.apache.hadoop.hdfs.protocol.DirectoryListing;
import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
import org.apache.hadoop.hdfs.protocol.HdfsConstants;
@ -68,9 +69,11 @@ import org.apache.hadoop.hdfs.server.blockmanagement.BlockManager;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockPlacementPolicy;
import org.apache.hadoop.hdfs.server.blockmanagement.BlockPlacementStatus;
import org.apache.hadoop.hdfs.server.blockmanagement.DatanodeDescriptor;
import org.apache.hadoop.hdfs.server.blockmanagement.DatanodeStorageInfo;
import org.apache.hadoop.hdfs.server.blockmanagement.NumberReplicas;
import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
import org.apache.hadoop.hdfs.server.datanode.CachingStrategy;
import org.apache.hadoop.hdfs.util.LightWeightLinkedSet;
import org.apache.hadoop.net.NetUtils;
import org.apache.hadoop.net.NetworkTopology;
import org.apache.hadoop.net.NodeBase;
@ -132,6 +135,8 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
private boolean showStoragePolcies = false;
private boolean showCorruptFileBlocks = false;
private boolean showReplicaDetails = false;
private long staleInterval;
/**
* True if we encountered an internal error during FSCK, such as not being
* able to delete a corrupt file.
@ -193,6 +198,9 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
networktopology,
namenode.getNamesystem().getBlockManager().getDatanodeManager()
.getHost2DatanodeMap());
this.staleInterval =
conf.getLong(DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_KEY,
DFSConfigKeys.DFS_NAMENODE_STALE_DATANODE_INTERVAL_DEFAULT);
for (Iterator<String> it = pmap.keySet().iterator(); it.hasNext();) {
String key = it.next();
@ -203,6 +211,9 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
else if (key.equals("blocks")) { this.showBlocks = true; }
else if (key.equals("locations")) { this.showLocations = true; }
else if (key.equals("racks")) { this.showRacks = true; }
else if (key.equals("replicadetails")) {
this.showReplicaDetails = true;
}
else if (key.equals("storagepolicies")) { this.showStoragePolcies = true; }
else if (key.equals("openforwrite")) {this.showOpenFiles = true; }
else if (key.equals("listcorruptfileblocks")) {
@ -502,9 +513,8 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
ExtendedBlock block = lBlk.getBlock();
boolean isCorrupt = lBlk.isCorrupt();
String blkName = block.toString();
DatanodeInfo[] locs = lBlk.getLocations();
NumberReplicas numberReplicas =
namenode.getNamesystem().getBlockManager().countNodes(block.getLocalBlock());
BlockManager bm = namenode.getNamesystem().getBlockManager();
NumberReplicas numberReplicas = bm.countNodes(block.getLocalBlock());
int liveReplicas = numberReplicas.liveReplicas();
int decommissionedReplicas = numberReplicas.decommissioned();;
int decommissioningReplicas = numberReplicas.decommissioning();
@ -513,6 +523,10 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
int totalReplicas = liveReplicas + decommissionedReplicas +
decommissioningReplicas;
res.totalReplicas += totalReplicas;
Collection<DatanodeDescriptor> corruptReplicas = null;
if (showReplicaDetails) {
corruptReplicas = bm.getCorruptReplicas(block.getLocalBlock());
}
short targetFileReplication = file.getReplication();
res.numExpectedReplicas += targetFileReplication;
if(totalReplicas < minReplication){
@ -573,14 +587,41 @@ public class NamenodeFsck implements DataEncryptionKeyFactory {
missize += block.getNumBytes();
} else {
report.append(" repl=" + liveReplicas);
if (showLocations || showRacks) {
if (showLocations || showRacks || showReplicaDetails) {
StringBuilder sb = new StringBuilder("[");
for (int j = 0; j < locs.length; j++) {
if (j > 0) { sb.append(", "); }
if (showRacks)
sb.append(NodeBase.getPath(locs[j]));
else
sb.append(locs[j]);
Iterable<DatanodeStorageInfo> storages = bm.getStorages(block.getLocalBlock());
for (Iterator<DatanodeStorageInfo> iterator = storages.iterator(); iterator.hasNext();) {
DatanodeStorageInfo storage = iterator.next();
DatanodeDescriptor dnDesc = storage.getDatanodeDescriptor();
if (showRacks) {
sb.append(NodeBase.getPath(dnDesc));
} else {
sb.append(new DatanodeInfoWithStorage(dnDesc, storage.getStorageID(), storage
.getStorageType()));
}
if (showReplicaDetails) {
LightWeightLinkedSet<Block> blocksExcess =
bm.excessReplicateMap.get(dnDesc.getDatanodeUuid());
sb.append("(");
if (dnDesc.isDecommissioned()) {
sb.append("DECOMMISSIONED)");
} else if (dnDesc.isDecommissionInProgress()) {
sb.append("DECOMMISSIONING)");
} else if (corruptReplicas != null && corruptReplicas.contains(dnDesc)) {
sb.append("CORRUPT)");
} else if (blocksExcess != null && blocksExcess.contains(block.getLocalBlock())) {
sb.append("EXCESS)");
} else if (dnDesc.isStale(this.staleInterval)) {
sb.append("STALE_NODE)");
} else if (storage.areBlockContentsStale()) {
sb.append("STALE_BLOCK_CONTENT)");
} else {
sb.append("LIVE)");
}
}
if (iterator.hasNext()) {
sb.append(", ");
}
}
sb.append(']');
report.append(" " + sb.toString());

6
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/tools/DFSck.java
View File

@ -95,7 +95,8 @@ public class DFSck extends Configured implements Tool {
+ "\t-storagepolicies\tprint out storage policy summary for the blocks\n\n"
+ "\t-blockId\tprint out which file this blockId belongs to, locations"
+ " (nodes, racks) of this block, and other diagnostics info"
+ " (under replicated, corrupted or not, etc)\n\n"
+ " (under replicated, corrupted or not, etc)\n"
+ "\t-replicaDetails\tprint out each replica details \n\n"
+ "Please Note:\n"
+ "\t1. By default fsck ignores files opened for write, "
+ "use -openforwrite to report such files. They are usually "
@ -266,6 +267,9 @@ public class DFSck extends Configured implements Tool {
else if (args[idx].equals("-blocks")) { url.append("&blocks=1"); }
else if (args[idx].equals("-locations")) { url.append("&locations=1"); }
else if (args[idx].equals("-racks")) { url.append("&racks=1"); }
else if (args[idx].equals("-replicaDetails")) {
url.append("&replicadetails=1");
}
else if (args[idx].equals("-storagepolicies")) { url.append("&storagepolicies=1"); }
else if (args[idx].equals("-list-corruptfileblocks")) {
url.append("&listcorruptfileblocks=1");

71
hadoop-hdfs-project/hadoop-hdfs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestFsck.java
View File

@ -779,6 +779,77 @@ public class TestFsck {
}
}
@Test(timeout = 60000)
public void testFsckReplicaDetails() throws Exception {
final short REPL_FACTOR = 1;
short NUM_DN = 1;
final long blockSize = 512;
final long fileSize = 1024;
boolean checkDecommissionInProgress = false;
String[] racks = { "/rack1" };
String[] hosts = { "host1" };
Configuration conf = new Configuration();
conf.setLong(DFSConfigKeys.DFS_BLOCK_SIZE_KEY, blockSize);
conf.setInt(DFSConfigKeys.DFS_REPLICATION_KEY, 1);
MiniDFSCluster cluster;
DistributedFileSystem dfs;
cluster =
new MiniDFSCluster.Builder(conf).numDataNodes(NUM_DN).hosts(hosts).racks(racks).build();
cluster.waitClusterUp();
dfs = cluster.getFileSystem();
// create files
final String testFile = new String("/testfile");
final Path path = new Path(testFile);
DFSTestUtil.createFile(dfs, path, fileSize, REPL_FACTOR, 1000L);
DFSTestUtil.waitReplication(dfs, path, REPL_FACTOR);
try {
// make sure datanode that has replica is fine before decommission
String fsckOut = runFsck(conf, 0, true, testFile, "-files", "-blocks", "-replicaDetails");
assertTrue(fsckOut.contains(NamenodeFsck.HEALTHY_STATUS));
assertTrue(fsckOut.contains("(LIVE)"));
// decommission datanode
ExtendedBlock eb = DFSTestUtil.getFirstBlock(dfs, path);
DatanodeDescriptor dn =
cluster.getNameNode().getNamesystem().getBlockManager()
.getBlockCollection(eb.getLocalBlock()).getBlocks()[0].getDatanode(0);
cluster.getNameNode().getNamesystem().getBlockManager().getDatanodeManager()
.getDecomManager().startDecommission(dn);
String dnName = dn.getXferAddr();
// check the replica status while decommissioning
fsckOut = runFsck(conf, 0, true, testFile, "-files", "-blocks", "-replicaDetails");
assertTrue(fsckOut.contains("(DECOMMISSIONING)"));
// Start 2nd Datanode and wait for decommission to start
cluster.startDataNodes(conf, 1, true, null, null, null);
DatanodeInfo datanodeInfo = null;
do {
Thread.sleep(2000);
for (DatanodeInfo info : dfs.getDataNodeStats()) {
if (dnName.equals(info.getXferAddr())) {
datanodeInfo = info;
}
}
if (!checkDecommissionInProgress && datanodeInfo != null
&& datanodeInfo.isDecommissionInProgress()) {
checkDecommissionInProgress = true;
}
} while (datanodeInfo != null && !datanodeInfo.isDecommissioned());
// check the replica status after decommission is done
fsckOut = runFsck(conf, 0, true, testFile, "-files", "-blocks", "-replicaDetails");
assertTrue(fsckOut.contains("(DECOMMISSIONED)"));
} finally {
if (cluster != null) {
cluster.shutdown();
}
}
}
/** Test if fsck can return -1 in case of failure
*