You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-commits@hadoop.apache.org by om...@apache.org on 2011/03/04 04:59:55 UTC
svn commit: r1077286 -
/hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java
Author: omalley
Date: Fri Mar 4 03:59:55 2011
New Revision: 1077286
URL: http://svn.apache.org/viewvc?rev=1077286&view=rev
Log:
commit 02e4668c2f4c571345080c7af0db424db581ee17
Author: Hairong Kuang <ha...@ucdev21.inktomisearch.com>
Date: Thu Mar 4 19:46:49 2010 +0000
HDFS:101 from http://issues.apache.org/jira/secure/attachment/12437819/detectDownDN3-0.20-yahoo.patch
+++ b/YAHOO-CHANGES.txt
+ HDFS-101. DFS write pipeline: DFSClient sometimes does not detect
+ second datanode failure. (hairong)
+
Modified:
hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java
Modified: hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java?rev=1077286&r1=1077285&r2=1077286&view=diff
==============================================================================
--- hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java (original)
+++ hadoop/common/branches/branch-0.20-security-patches/src/hdfs/org/apache/hadoop/hdfs/server/datanode/BlockReceiver.java Fri Mar 4 03:59:55 2011
@@ -75,6 +75,7 @@ class BlockReceiver implements java.io.C
DatanodeInfo srcDataNode = null;
private Checksum partialCrc = null;
private DataNode datanode = null;
+ volatile private boolean mirrorError;
BlockReceiver(Block block, DataInputStream in, String inAddr,
String myAddr, boolean isRecovery, String clientName,
@@ -176,21 +177,19 @@ class BlockReceiver implements java.io.C
/**
* While writing to mirrorOut, failure to write to mirror should not
- * affect this datanode unless a client is writing the block.
+ * affect this datanode.
*/
private void handleMirrorOutError(IOException ioe) throws IOException {
- LOG.info(datanode.dnRegistration + ":Exception writing block " +
+ LOG.info(datanode.dnRegistration + ": Exception writing block " +
block + " to mirror " + mirrorAddr + "\n" +
StringUtils.stringifyException(ioe));
- mirrorOut = null;
- //
- // If stream-copy fails, continue
- // writing to disk for replication requests. For client
- // writes, return error so that the client can do error
- // recovery.
- //
- if (clientName.length() > 0) {
+ if (Thread.interrupted()) { // shut down if the thread is interrupted
throw ioe;
+ } else { // encounter an error while writing to mirror
+ // continue to run even if can not write to mirror
+ // notify client of the error
+ // and wait for the client to shut down the pipeline
+ mirrorError = true;
}
}
@@ -399,8 +398,8 @@ class BlockReceiver implements java.io.C
setBlockPosition(offsetInBlock);
- //First write the packet to the mirror:
- if (mirrorOut != null) {
+ // First write the packet to the mirror:
+ if (mirrorOut != null && !mirrorError) {
try {
mirrorOut.write(buf.array(), buf.position(), buf.remaining());
mirrorOut.flush();
@@ -518,7 +517,8 @@ class BlockReceiver implements java.io.C
if (clientName.length() > 0) {
responder = new Daemon(datanode.threadGroup,
new PacketResponder(this, block, mirrIn,
- replyOut, numTargets));
+ replyOut, numTargets,
+ Thread.currentThread()));
responder.start(); // start thread to processes reponses
}
@@ -703,18 +703,21 @@ class BlockReceiver implements java.io.C
DataOutputStream replyOut; // output to upstream datanode
private int numTargets; // number of downstream datanodes including myself
private BlockReceiver receiver; // The owner of this responder.
+ private Thread receiverThread; // the thread that spawns this responder
public String toString() {
return "PacketResponder " + numTargets + " for Block " + this.block;
}
PacketResponder(BlockReceiver receiver, Block b, DataInputStream in,
- DataOutputStream out, int numTargets) {
+ DataOutputStream out, int numTargets,
+ Thread receiverThread) {
this.receiver = receiver;
this.block = b;
mirrorIn = in;
replyOut = out;
this.numTargets = numTargets;
+ this.receiverThread = receiverThread;
}
/**
@@ -861,25 +864,27 @@ class BlockReceiver implements java.io.C
}
boolean lastPacketInBlock = false;
+ boolean isInterrupted = false;
final long startTime = ClientTraceLog.isInfoEnabled() ? System.nanoTime() : 0;
while (running && datanode.shouldRun && !lastPacketInBlock) {
try {
- boolean didRead = false;
long expected = -2;
PipelineAck ack = new PipelineAck();
+ long seqno = -2;
try {
- // read an ack from downstream datanode
- ack.readFields(mirrorIn);
- if (LOG.isDebugEnabled()) {
- LOG.debug("PacketResponder " + numTargets + " got " + ack);
- }
- long seqno = ack.getSeqno();
- didRead = true;
+ if (!mirrorError) {
+ // read an ack from downstream datanode
+ ack.readFields(mirrorIn);
+ if (LOG.isDebugEnabled()) {
+ LOG.debug("PacketResponder " + numTargets + " got " + ack);
+ }
+ seqno = ack.getSeqno();
+ }
if (seqno == PipelineAck.HEART_BEAT.getSeqno()) {
ack.write(replyOut); // send keepalive
replyOut.flush();
- } else if (seqno >= 0) {
+ } else if (seqno >= 0 || mirrorError) {
Packet pkt = null;
synchronized (this) {
while (running && datanode.shouldRun && ackQueue.size() == 0) {
@@ -891,10 +896,13 @@ class BlockReceiver implements java.io.C
}
wait();
}
+ if (!running || !datanode.shouldRun) {
+ break;
+ }
pkt = ackQueue.removeFirst();
expected = pkt.seqno;
notifyAll();
- if (seqno != expected) {
+ if (seqno != expected && !mirrorError) {
throw new IOException("PacketResponder " + numTargets +
" for block " + block +
" expected seqno:" + expected +
@@ -903,27 +911,32 @@ class BlockReceiver implements java.io.C
lastPacketInBlock = pkt.lastPacketInBlock;
}
}
- } catch (Throwable e) {
- if (running) {
- LOG.info("PacketResponder " + block + " " + numTargets +
- " Exception " + StringUtils.stringifyException(e));
- running = false;
+ } catch (InterruptedException ine) {
+ isInterrupted = true;
+ } catch (IOException ioe) {
+ if (Thread.interrupted()) {
+ isInterrupted = true;
+ } else {
+ // continue to run even if can not read from mirror
+ // notify client of the error
+ // and wait for the client to shut down the pipeline
+ mirrorError = true;
+ LOG.info("PacketResponder " + block + " " + numTargets +
+ " Exception " + StringUtils.stringifyException(ioe));
}
}
- if (Thread.interrupted()) {
+ if (Thread.interrupted() || isInterrupted) {
/* The receiver thread cancelled this thread.
* We could also check any other status updates from the
* receiver thread (e.g. if it is ok to write to replyOut).
* It is prudent to not send any more status back to the client
* because this datanode has a problem. The upstream datanode
- * will detect a timout on heartbeats and will declare that
- * this datanode is bad, and rightly so.
+ * will detect that this datanode is bad, and rightly so.
*/
LOG.info("PacketResponder " + block + " " + numTargets +
" : Thread is interrupted.");
- running = false;
- continue;
+ break;
}
// If this is the last packet in block, then close block
@@ -952,7 +965,7 @@ class BlockReceiver implements java.io.C
// construct my ack message
short[] replies = null;
- if (!didRead) { // no ack is read
+ if (mirrorError) { // no ack is read
replies = new short[2];
replies[0] = DataTransferProtocol.OP_STATUS_SUCCESS;
replies[1] = DataTransferProtocol.OP_STATUS_ERROR;
@@ -973,30 +986,15 @@ class BlockReceiver implements java.io.C
" for block " + block +
" responded an ack: " + replyAck);
}
-
- // If we forwarded an error response from a downstream datanode
- // and we are acting on behalf of a client, then we quit. The
- // client will drive the recovery mechanism.
- if (!replyAck.isSuccess() && receiver.clientName.length() > 0) {
- running = false;
- }
- } catch (IOException e) {
+ } catch (Throwable e) {
LOG.warn("IOException in BlockReceiver.run(): ", e);
if (running) {
- try {
- datanode.checkDiskError(e); // may throw an exception here
- } catch (IOException ioe) {
- LOG.warn("DataNode.chekDiskError failed in run() with: ", ioe);
- }
LOG.info("PacketResponder " + block + " " + numTargets +
" Exception " + StringUtils.stringifyException(e));
running = false;
}
- } catch (RuntimeException e) {
- if (running) {
- LOG.info("PacketResponder " + block + " " + numTargets +
- " Exception " + StringUtils.stringifyException(e));
- running = false;
+ if (!Thread.interrupted()) { // error not caused by interruption
+ receiverThread.interrupt();
}
}
}