HDFS-9034. StorageTypeStats Metric should not count failed storage. Contributed by Surendra Singh Lilhore.
This commit is contained in:
parent
867048c3e4
commit
df83230948
@ -963,6 +963,7 @@ nodes with its data cleared (or user can just remove the StorageID
|
||||
// no need to update its timestamp
|
||||
// because its is done when the descriptor is created
|
||||
heartbeatManager.addDatanode(nodeDescr);
|
||||
heartbeatManager.updateDnStat(nodeDescr);
|
||||
incrementVersionCount(nodeReg.getSoftwareVersion());
|
||||
startDecommissioningIfExcluded(nodeDescr);
|
||||
success = true;
|
||||
|
@ -19,6 +19,7 @@
|
||||
|
||||
import org.apache.hadoop.fs.StorageType;
|
||||
import org.apache.hadoop.hdfs.DFSUtilClient;
|
||||
import org.apache.hadoop.hdfs.server.protocol.DatanodeStorage;
|
||||
|
||||
import java.util.EnumMap;
|
||||
import java.util.HashSet;
|
||||
@ -61,8 +62,10 @@ synchronized void add(final DatanodeDescriptor node) {
|
||||
}
|
||||
Set<StorageType> storageTypes = new HashSet<>();
|
||||
for (DatanodeStorageInfo storageInfo : node.getStorageInfos()) {
|
||||
statsMap.addStorage(storageInfo, node);
|
||||
storageTypes.add(storageInfo.getStorageType());
|
||||
if (storageInfo.getState() != DatanodeStorage.State.FAILED) {
|
||||
statsMap.addStorage(storageInfo, node);
|
||||
storageTypes.add(storageInfo.getStorageType());
|
||||
}
|
||||
}
|
||||
for (StorageType storageType : storageTypes) {
|
||||
statsMap.addNode(storageType, node);
|
||||
@ -86,8 +89,10 @@ synchronized void subtract(final DatanodeDescriptor node) {
|
||||
}
|
||||
Set<StorageType> storageTypes = new HashSet<>();
|
||||
for (DatanodeStorageInfo storageInfo : node.getStorageInfos()) {
|
||||
statsMap.subtractStorage(storageInfo, node);
|
||||
storageTypes.add(storageInfo.getStorageType());
|
||||
if (storageInfo.getState() != DatanodeStorage.State.FAILED) {
|
||||
statsMap.subtractStorage(storageInfo, node);
|
||||
storageTypes.add(storageInfo.getStorageType());
|
||||
}
|
||||
}
|
||||
for (StorageType storageType : storageTypes) {
|
||||
statsMap.subtractNode(storageType, node);
|
||||
@ -202,10 +207,12 @@ private void subtractStorage(final DatanodeStorageInfo info,
|
||||
|
||||
private void subtractNode(StorageType storageType,
|
||||
final DatanodeDescriptor node) {
|
||||
StorageTypeStats storageTypeStats =
|
||||
storageTypeStatsMap.get(storageType);
|
||||
StorageTypeStats storageTypeStats = storageTypeStatsMap.get(storageType);
|
||||
if (storageTypeStats != null) {
|
||||
storageTypeStats.subtractNode(node);
|
||||
if (storageTypeStats.getNodesInService() == 0) {
|
||||
storageTypeStatsMap.remove(storageType);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
@ -203,6 +203,7 @@ synchronized void register(final DatanodeDescriptor d) {
|
||||
|
||||
//update its timestamp
|
||||
d.updateHeartbeatState(StorageReport.EMPTY_ARRAY, 0L, 0L, 0, 0, null);
|
||||
stats.add(d);
|
||||
}
|
||||
}
|
||||
|
||||
@ -212,11 +213,14 @@ synchronized DatanodeDescriptor[] getDatanodes() {
|
||||
|
||||
synchronized void addDatanode(final DatanodeDescriptor d) {
|
||||
// update in-service node count
|
||||
stats.add(d);
|
||||
datanodes.add(d);
|
||||
d.setAlive(true);
|
||||
}
|
||||
|
||||
void updateDnStat(final DatanodeDescriptor d){
|
||||
stats.add(d);
|
||||
}
|
||||
|
||||
synchronized void removeDatanode(DatanodeDescriptor node) {
|
||||
if (node.isAlive()) {
|
||||
stats.subtract(node);
|
||||
|
@ -18,10 +18,12 @@
|
||||
package org.apache.hadoop.hdfs.server.blockmanagement;
|
||||
|
||||
import static org.junit.Assert.assertEquals;
|
||||
import static org.junit.Assert.assertFalse;
|
||||
import static org.junit.Assert.assertNotNull;
|
||||
import static org.junit.Assert.assertTrue;
|
||||
import static org.junit.Assert.fail;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.IOException;
|
||||
import java.io.InputStream;
|
||||
import java.net.URL;
|
||||
@ -29,9 +31,12 @@
|
||||
import java.util.Map;
|
||||
import java.util.Set;
|
||||
|
||||
import org.apache.hadoop.fs.Path;
|
||||
import org.apache.hadoop.fs.StorageType;
|
||||
import org.apache.hadoop.hdfs.DFSTestUtil;
|
||||
import org.apache.hadoop.hdfs.HdfsConfiguration;
|
||||
import org.apache.hadoop.hdfs.MiniDFSCluster;
|
||||
import org.apache.hadoop.hdfs.server.datanode.DataNodeTestUtils;
|
||||
import org.junit.After;
|
||||
import org.junit.Before;
|
||||
import org.junit.Test;
|
||||
@ -143,4 +148,61 @@ public void testStorageTypeStatsJMX() throws Exception {
|
||||
assertTrue(typesPresent.contains("DISK"));
|
||||
assertTrue(typesPresent.contains("RAM_DISK"));
|
||||
}
|
||||
|
||||
@Test
|
||||
public void testStorageTypeStatsWhenStorageFailed() throws Exception {
|
||||
DFSTestUtil.createFile(cluster.getFileSystem(),
|
||||
new Path("/blockStatsFile1"), 1024, (short) 1, 0L);
|
||||
Map<StorageType, StorageTypeStats> storageTypeStatsMap = cluster
|
||||
.getNamesystem().getBlockManager().getStorageTypeStats();
|
||||
|
||||
StorageTypeStats storageTypeStats = storageTypeStatsMap
|
||||
.get(StorageType.RAM_DISK);
|
||||
assertEquals(6, storageTypeStats.getNodesInService());
|
||||
|
||||
storageTypeStats = storageTypeStatsMap.get(StorageType.DISK);
|
||||
assertEquals(3, storageTypeStats.getNodesInService());
|
||||
|
||||
storageTypeStats = storageTypeStatsMap.get(StorageType.ARCHIVE);
|
||||
assertEquals(3, storageTypeStats.getNodesInService());
|
||||
String dataDir = cluster.getDataDirectory();
|
||||
File dn1ArcVol1 = new File(dataDir, "data" + (3 * 0 + 2));
|
||||
File dn2ArcVol1 = new File(dataDir, "data" + (3 * 1 + 2));
|
||||
File dn3ArcVol1 = new File(dataDir, "data" + (3 * 2 + 2));
|
||||
DataNodeTestUtils.injectDataDirFailure(dn1ArcVol1);
|
||||
DataNodeTestUtils.injectDataDirFailure(dn2ArcVol1);
|
||||
DataNodeTestUtils.injectDataDirFailure(dn3ArcVol1);
|
||||
try {
|
||||
DFSTestUtil.createFile(cluster.getFileSystem(), new Path(
|
||||
"/blockStatsFile2"), 1024, (short) 1, 0L);
|
||||
fail("Should throw exception, becuase no DISK storage available");
|
||||
} catch (Exception e) {
|
||||
assertTrue(e.getMessage().contains(
|
||||
"could only be replicated to 0 nodes instead"));
|
||||
}
|
||||
// wait for heartbeat
|
||||
Thread.sleep(6000);
|
||||
storageTypeStatsMap = cluster.getNamesystem().getBlockManager()
|
||||
.getStorageTypeStats();
|
||||
assertFalse("StorageTypeStatsMap should not contain DISK Storage type",
|
||||
storageTypeStatsMap.containsKey(StorageType.DISK));
|
||||
DataNodeTestUtils.restoreDataDirFromFailure(dn1ArcVol1);
|
||||
DataNodeTestUtils.restoreDataDirFromFailure(dn2ArcVol1);
|
||||
DataNodeTestUtils.restoreDataDirFromFailure(dn3ArcVol1);
|
||||
for (int i = 0; i < 3; i++) {
|
||||
cluster.restartDataNode(0, true);
|
||||
}
|
||||
// wait for heartbeat
|
||||
Thread.sleep(6000);
|
||||
storageTypeStatsMap = cluster.getNamesystem().getBlockManager()
|
||||
.getStorageTypeStats();
|
||||
storageTypeStats = storageTypeStatsMap.get(StorageType.RAM_DISK);
|
||||
assertEquals(6, storageTypeStats.getNodesInService());
|
||||
|
||||
storageTypeStats = storageTypeStatsMap.get(StorageType.DISK);
|
||||
assertEquals(3, storageTypeStats.getNodesInService());
|
||||
|
||||
storageTypeStats = storageTypeStatsMap.get(StorageType.ARCHIVE);
|
||||
assertEquals(3, storageTypeStats.getNodesInService());
|
||||
}
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user