HDFS-12914. Block report leases cause missing blocks until next report. Contributed by Santosh Marella, He Xiaoqiao.

Signed-off-by: Wei-Chiu Chuang <weichiu@apache.org>
Co-authored-by: He Xiaoqiao <hexiaoqiao@apache.org>
This commit is contained in:
Santosh Marella 2019-06-14 10:35:33 -07:00 committed by Wei-Chiu Chuang
parent 3ba090f436
commit ae4143a529
2 changed files with 34 additions and 21 deletions

View File

@ -2572,6 +2572,21 @@ public class BlockManager implements BlockStatsMXBean {
}
}
/**
* Check block report lease.
* @return true if lease exist and not expire
*/
public boolean checkBlockReportLease(BlockReportContext context,
final DatanodeID nodeID) throws UnregisteredNodeException {
if (context == null) {
return true;
}
DatanodeDescriptor node = datanodeManager.getDatanode(nodeID);
final long startTime = Time.monotonicNow();
return blockReportLeaseManager.checkLease(node, startTime,
context.getLeaseId());
}
/**
* The given storage is reporting all its blocks.
* Update the (storage{@literal -->}block list) and
@ -2619,12 +2634,6 @@ public class BlockManager implements BlockStatsMXBean {
blockReportLeaseManager.removeLease(node);
return !node.hasStaleStorages();
}
if (context != null) {
if (!blockReportLeaseManager.checkLease(node, startTime,
context.getLeaseId())) {
return false;
}
}
if (storageInfo.getBlockReportCount() == 0) {
// The first block report can be processed a lot more efficiently than

View File

@ -45,7 +45,6 @@ import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.concurrent.Callable;
import com.google.common.collect.Lists;
@ -175,6 +174,7 @@ import org.apache.hadoop.hdfs.server.protocol.NamenodeProtocols;
import org.apache.hadoop.hdfs.server.protocol.NamenodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.NamespaceInfo;
import org.apache.hadoop.hdfs.server.protocol.NodeRegistration;
import org.apache.hadoop.hdfs.server.protocol.RegisterCommand;
import org.apache.hadoop.hdfs.server.protocol.RemoteEditLogManifest;
import org.apache.hadoop.hdfs.server.protocol.SlowDiskReports;
import org.apache.hadoop.hdfs.server.protocol.SlowPeerReports;
@ -1591,6 +1591,8 @@ public class NameNodeRpcServer implements NamenodeProtocols {
}
final BlockManager bm = namesystem.getBlockManager();
boolean noStaleStorages = false;
try {
if (bm.checkBlockReportLease(context, nodeReg)) {
for (int r = 0; r < reports.length; r++) {
final BlockListAsLongs blocks = reports[r].getBlocks();
//
@ -1599,13 +1601,15 @@ public class NameNodeRpcServer implements NamenodeProtocols {
// call of this loop is the final updated value for noStaleStorage.
//
final int index = r;
noStaleStorages = bm.runBlockOp(new Callable<Boolean>() {
@Override
public Boolean call() throws IOException {
return bm.processReport(nodeReg, reports[index].getStorage(),
blocks, context);
noStaleStorages = bm.runBlockOp(() ->
bm.processReport(nodeReg, reports[index].getStorage(),
blocks, context));
}
});
}
} catch (UnregisteredNodeException une) {
LOG.debug("Datanode {} is attempting to report but not register yet.",
nodeReg);
return RegisterCommand.REGISTER;
}
bm.removeBRLeaseIfNeeded(nodeReg, context);