Skip to content

HDFS-17052. Improve BlockPlacementPolicyRackFaultTolerant to avoid choose nodes failed when no enough Rack. #5759

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 2, 2023
Merged

HDFS-17052. Improve BlockPlacementPolicyRackFaultTolerant to avoid choose nodes failed when no enough Rack. #5759

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 14 additions & 3 deletions .../org/apache/hadoop/hdfs/server/blockmanagement/BlockPlacementPolicyRackFaultTolerant.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -170,7 +170,7 @@ private void chooseEvenlyFromRemainingRacks(Node writer,
NotEnoughReplicasException lastException = e;
int bestEffortMaxNodesPerRack = maxNodesPerRack;
while (results.size() != totalReplicaExpected &&
numResultsOflastChoose != results.size()) {
bestEffortMaxNodesPerRack < totalReplicaExpected) {
// Exclude the chosen nodes
final Set<Node> newExcludeNodes = new HashSet<>();
for (DatanodeStorageInfo resultStorage : results) {
Expand All @@ -192,11 +192,22 @@ private void chooseEvenlyFromRemainingRacks(Node writer,
} finally {
excludedNodes.addAll(newExcludeNodes);
}
// To improve performance, the maximum value of 'bestEffortMaxNodesPerRack'
// is calculated only when it is not possible to select a node.
if (numResultsOflastChoose == results.size()) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It will be better to add some annotation here.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for your suggestion, I'll make the required code changes promptly

Map<String, Integer> nodesPerRack = new HashMap<>();
for (DatanodeStorageInfo dsInfo : results) {
String rackName = dsInfo.getDatanodeDescriptor().getNetworkLocation();
nodesPerRack.merge(rackName, 1, Integer::sum);
}
bestEffortMaxNodesPerRack =
Math.max(bestEffortMaxNodesPerRack, Collections.max(nodesPerRack.values()));
}
}

if (numResultsOflastChoose != totalReplicaExpected) {
if (results.size() != totalReplicaExpected) {
LOG.debug("Best effort placement failed: expecting {} replicas, only "
+ "chose {}.", totalReplicaExpected, numResultsOflastChoose);
+ "chose {}.", totalReplicaExpected, results.size());
throw lastException;
}
}
Expand Down
61 changes: 61 additions & 0 deletions ...fs/src/test/java/org/apache/hadoop/hdfs/server/namenode/TestReconstructStripedBlocks.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.StorageType;
import org.apache.hadoop.hdfs.DFSConfigKeys;
import org.apache.hadoop.hdfs.DFSTestUtil;
import org.apache.hadoop.hdfs.DistributedFileSystem;
Expand Down Expand Up @@ -52,6 +53,7 @@
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.util.Arrays;
import java.util.BitSet;
import java.util.Iterator;
import java.util.List;
Expand Down Expand Up @@ -515,4 +517,63 @@ public void testReconstrutionWithBusyBlock1() throws Exception {
assertEquals(9, bm.countNodes(blockInfo).liveReplicas());
}

@Test
public void testReconstructionWithStorageTypeNotEnough() throws Exception {
final HdfsConfiguration conf = new HdfsConfiguration();
conf.setInt(DFSConfigKeys.DFS_NAMENODE_REDUNDANCY_INTERVAL_SECONDS_KEY, 1);

// Nine disk node eleven archive node.
int numDn = groupSize * 2 + 2;
StorageType[][] storageTypes = new StorageType[numDn][];
Arrays.fill(storageTypes, 0, groupSize,
new StorageType[]{StorageType.DISK, StorageType.DISK});
Arrays.fill(storageTypes, groupSize, numDn,
new StorageType[]{StorageType.ARCHIVE, StorageType.ARCHIVE});

// Nine disk racks and one archive rack.
String[] racks = {
"/rack1", "/rack2", "/rack3", "/rack4", "/rack5", "/rack6", "/rack7", "/rack8",
"/rack9", "/rack0", "/rack0", "/rack0", "/rack0", "/rack0", "/rack0", "/rack0",
"/rack0", "/rack0", "/rack0", "/rack0"};

cluster = new MiniDFSCluster.Builder(conf).numDataNodes(numDn)
.storageTypes(storageTypes)
.racks(racks)
.build();
cluster.waitActive();
DistributedFileSystem fs = cluster.getFileSystem();
fs.enableErasureCodingPolicy(
StripedFileTestUtil.getDefaultECPolicy().getName());

try {
fs.mkdirs(dirPath);
fs.setStoragePolicy(dirPath, "COLD");
fs.setErasureCodingPolicy(dirPath,
StripedFileTestUtil.getDefaultECPolicy().getName());
DFSTestUtil.createFile(fs, filePath,
cellSize * dataBlocks * 2, (short) 1, 0L);

// Stop one dn.
LocatedBlocks blks = fs.getClient().getLocatedBlocks(filePath.toString(), 0);
LocatedStripedBlock block = (LocatedStripedBlock) blks.getLastLocatedBlock();
DatanodeInfo dnToStop = block.getLocations()[0];
cluster.stopDataNode(dnToStop.getXferAddr());
cluster.setDataNodeDead(dnToStop);

// Wait for reconstruction to happen.
StripedFileTestUtil.waitForReconstructionFinished(filePath, fs, groupSize);
blks = fs.getClient().getLocatedBlocks(filePath.toString(), 0);
block = (LocatedStripedBlock) blks.getLastLocatedBlock();
BitSet bitSet = new BitSet(groupSize);
for (byte index : block.getBlockIndices()) {
bitSet.set(index);
}
for (int i = 0; i < groupSize; i++) {
Assert.assertTrue(bitSet.get(i));
}
} finally {
cluster.shutdown();
}
}

}