Skip to content

HDFS-14318:dn cannot be recognized and must be restarted to recognize the Repaired disk #1104

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

Open
wants to merge 5 commits into
base: trunk
Choose a base branch
from
+91 −10
Open

HDFS-14318:dn cannot be recognized and must be restarted to recognize the Repaired disk #1104

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
ec5c73c
HDFS-14318:dn cannot be recognized and must be restarted to recognize…
hunshenshi Jul 16, 2019
09c21ac
Merge branch 'trunk' into HDFS-14318
hunshenshi Jul 16, 2019
f8469f9
Merge branch 'trunk' into HDFS-14318
hunshenshi Sep 15, 2019
e53e149
use ScheduledExecutorService and fix some bugs
hunshenshi Sep 15, 2019
919df08
Merge branch 'trunk' into HDFS-14318
hunshenshi Oct 21, 2019
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
4 changes: 4 additions & 0 deletions hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -732,6 +732,10 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
public static final String DFS_DATANODE_DISK_CHECK_TIMEOUT_DEFAULT =
"10m";

public static final String DFS_DATANODE_DISK_CHECK_INTERVAL_KEY =
"dfs.datanode.disk.check.interval";
public static final int DFS_DATANODE_DISK_CHECK_INTERVAL_DEFAULT = 5000;

public static final String DFS_NAMENODE_EC_POLICIES_MAX_CELLSIZE_KEY = "dfs.namenode.ec.policies.max.cellsize";
public static final int DFS_NAMENODE_EC_POLICIES_MAX_CELLSIZE_DEFAULT = 4 * 1024 * 1024;
public static final String DFS_NAMENODE_EC_POLICIES_USERPOLICIES_ALLOWED_KEY =
Expand Down
97 changes: 87 additions & 10 deletions ...fs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,9 +46,14 @@
import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_MAX_NUM_BLOCKS_TO_LOG_KEY;
import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_METRICS_LOGGER_PERIOD_SECONDS_DEFAULT;
import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_METRICS_LOGGER_PERIOD_SECONDS_KEY;

import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_DISK_CHECK_INTERVAL_KEY;
import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_DISK_CHECK_INTERVAL_DEFAULT;

import static org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage.PIPELINE_SETUP_APPEND_RECOVERY;
import static org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage.PIPELINE_SETUP_CREATE;
import static org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage.PIPELINE_SETUP_STREAMING_RECOVERY;

import static org.apache.hadoop.util.ExitUtil.terminate;

import org.apache.hadoop.fs.CommonConfigurationKeysPublic;
Expand Down Expand Up @@ -84,13 +89,7 @@
import java.util.Map.Entry;
import java.util.Set;
import java.util.UUID;
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.ScheduledThreadPoolExecutor;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.*;
import java.util.concurrent.atomic.AtomicInteger;

import javax.annotation.Nullable;
Expand Down Expand Up @@ -407,6 +406,10 @@ public static InetSocketAddress createSocketAddr(String target) {
private final StorageLocationChecker storageLocationChecker;

private final DatasetVolumeChecker volumeChecker;
volatile FsDatasetSpi<? extends FsVolumeSpi> allData = null;
private ScheduledExecutorService scheduledExecutor;
private int checkDiskInterval = 5*1000;
private List<StorageLocation> errorDisk = null;
Copy link
Contributor

Choose a reason for hiding this comment

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

allData = data + errorDisk right? It seems that allData is redundant variable.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

allData = data + errorDisk is right. allData will use in volumeChecker.checkAllVolumes(allData).
If no allData, we must make a variable before volumeChecker.checkAllVolumes, the variable is data + errorDisk. So I think allData is ok .


private final SocketFactory socketFactory;

Expand Down Expand Up @@ -445,6 +448,8 @@ private static Tracer createTracer(Configuration conf) {
volumeChecker = new DatasetVolumeChecker(conf, new Timer());
this.xferService =
HadoopExecutors.newCachedThreadPool(new Daemon.DaemonFactory());
this.checkDiskInterval = conf.getInt(DFS_DATANODE_DISK_CHECK_INTERVAL_KEY,
DFS_DATANODE_DISK_CHECK_INTERVAL_DEFAULT);
}

/**
Expand Down Expand Up @@ -478,6 +483,8 @@ private static Tracer createTracer(Configuration conf) {
this.pipelineSupportECN = conf.getBoolean(
DFSConfigKeys.DFS_PIPELINE_ECN_ENABLED,
DFSConfigKeys.DFS_PIPELINE_ECN_ENABLED_DEFAULT);
this.checkDiskInterval = conf.getInt(DFS_DATANODE_DISK_CHECK_INTERVAL_KEY,
DFS_DATANODE_DISK_CHECK_INTERVAL_DEFAULT);

confVersion = "core-" +
conf.get("hadoop.common.configuration.version", "UNSPECIFIED") +
Expand Down Expand Up @@ -801,6 +808,7 @@ private synchronized void refreshVolumes(String newVolumes) throws IOException {
public IOException call() {
try {
data.addVolume(location, nsInfos);
allData.addVolume(location, nsInfos);
} catch (IOException e) {
return e;
}
Expand All @@ -822,6 +830,14 @@ public IOException call() {
} else {
effectiveVolumes.add(volume.toString());
LOG.info("Successfully added volume: {}", volume);
if (errorDisk != null && !errorDisk.isEmpty()) {
Copy link
Contributor

Choose a reason for hiding this comment

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

It may be not necessary to check if errorDisk is empty.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I don't think so. If refreshVolumes is invoked by reconfigurePropertyImpl, errorDisk may be null.

  public String reconfigurePropertyImpl(String property, String newVal)
      throws ReconfigurationException {
    switch (property) {
      case DFS_DATANODE_DATA_DIR_KEY: {
        IOException rootException = null;
        try {
          LOG.info("Reconfiguring {} to {}", property, newVal);
          this.refreshVolumes(newVal);
...
  }

LOG.debug("check errorDisk for {} disk ", volume);
if (errorDisk.contains(volume)) {
errorDisk.remove(volume);
LOG.info("Remove {} from errorDisk, " +
"because of the repaired disk ", volume);
}
}
}
} catch (Exception e) {
errorMessageBuilder.append(
Expand Down Expand Up @@ -1701,6 +1717,13 @@ void initBlockPool(BPOfferService bpos) throws IOException {
// Exclude failed disks before initializing the block pools to avoid startup
// failures.
checkDiskError();

// start check disk thread.
scheduledExecutor = Executors.newScheduledThreadPool(1);
Runnable checkDisk = new CheckDisk();
scheduledExecutor.scheduleAtFixedRate(checkDisk, checkDiskInterval,
checkDiskInterval, TimeUnit.SECONDS);

try {
data.addBlockPool(nsInfo.getBlockPoolID(), getConf());
} catch (AddBlockPoolException e) {
Expand Down Expand Up @@ -1775,6 +1798,7 @@ private void initStorage(final NamespaceInfo nsInfo) throws IOException {
synchronized(this) {
if (data == null) {
data = factory.newInstance(this, storage, getConf());
allData = factory.newInstance(this, storage, getConf());
}
}
}
Expand Down Expand Up @@ -2194,6 +2218,26 @@ public void shutdown() {
tracer.close();
}

private class CheckDisk implements Runnable {

@Override
public void run() {
while (shouldRun) {
LOG.info("CheckDiskThread running ");
if (errorDisk != null && !errorDisk.isEmpty()) {
try {
checkDiskError();
} catch (Exception e) {
LOG.warn("Unexpected exception occurred while" +
" checking disk error "+ e);
return;
}
lastDiskErrorCheck = Time.monotonicNow();
}
}
}
}

/**
* Check if there is a disk failure asynchronously
* and if so, handle the error.
Expand Down Expand Up @@ -3422,20 +3466,53 @@ public ShortCircuitRegistry getShortCircuitRegistry() {
@VisibleForTesting
public void checkDiskError() throws IOException {
Set<FsVolumeSpi> unhealthyVolumes;
Configuration conf = getConf();
String newDataDirs = null;
try {
unhealthyVolumes = volumeChecker.checkAllVolumes(data);
// check all volume
unhealthyVolumes = volumeChecker.checkAllVolumes(allData);
lastDiskErrorCheck = Time.monotonicNow();
} catch (InterruptedException e) {
LOG.error("Interruped while running disk check", e);
LOG.error("Interrupted while running disk check", e);
throw new IOException("Interrupted while running disk check", e);
}

if (unhealthyVolumes.size() > 0) {
if (errorDisk == null) {
errorDisk = new ArrayList<>();
}
List<StorageLocation> tmpDisk = Lists.newArrayList(errorDisk);
errorDisk.clear();
for (FsVolumeSpi vol : unhealthyVolumes) {
LOG.info("Add error disk {} to errorDisk - {}",
vol.getStorageLocation(), vol.getStorageLocation());
errorDisk.add(vol.getStorageLocation());
if (tmpDisk.contains(vol.getStorageLocation())) {
tmpDisk.remove(vol.getStorageLocation());
}
}
LOG.warn("checkDiskError got {} failed volumes - {}",
unhealthyVolumes.size(), unhealthyVolumes);
handleVolumeFailures(unhealthyVolumes);
if (!tmpDisk.isEmpty()) {
newDataDirs = conf.get(DFS_DATANODE_DATA_DIR_KEY)
+ "," + Joiner.on(",").join(tmpDisk);
}
} else {
LOG.debug("checkDiskError encountered no failures");
LOG.debug("checkDiskError encountered no failures," +
"then check errorDisk");
if (errorDisk != null && !errorDisk.isEmpty()) {
newDataDirs = conf.get(DFS_DATANODE_DATA_DIR_KEY)
+ "," + Joiner.on(",").join(errorDisk);
}
}
if (newDataDirs != null) {
LOG.debug("Bad disks is repaired, should refreshVolumes disk.");
try {
refreshVolumes(newDataDirs);
} catch (IOException e) {
LOG.error("Bad disks is repaired, refreshVolumes error : ", e);
}
}
}

Expand Down