[
https://issues.apache.org/jira/browse/HDFS-16757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17905954#comment-17905954
]
ASF GitHub Bot commented on HDFS-16757:
---------------------------------------
tomscut commented on code in PR #6926:
URL: https://github.com/apache/hadoop/pull/6926#discussion_r1886463808
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/fsdataset/impl/BlockPoolSlice.java:
##########
@@ -1152,4 +1153,14 @@ void setDeleteDuplicateReplicasForTests(
this.deleteDuplicateReplicas = deleteDuplicateReplicasForTests;
}
+ public File hardLinkOneBlock(File src, File srcMeta, Block dstBlock) throws
IOException {
+ File dstMeta = new File(tmpDir,
+ DatanodeUtil.getMetaName(dstBlock.getBlockName(),
dstBlock.getGenerationStamp()));
+ HardLink.createHardLink(srcMeta, dstMeta);
+
+ File dstBlockFile = new File(tmpDir, dstBlock.getBlockName());
+ HardLink.createHardLink(src, dstBlockFile);
Review Comment:
Hi @LiuGuH @haiyang1987 , do you use this feature on a large scale? Does
creating a lot of hard links have a maintenance impact on the system? Have you
considered implementing it based on rename?
##########
hadoop-hdfs-project/hadoop-hdfs-client/src/main/java/org/apache/hadoop/hdfs/DFSClient.java:
##########
@@ -1956,6 +1957,26 @@ protected IOStreamPair connectToDN(DatanodeInfo dn, int
timeout,
socketFactory, getConf().isConnectToDnViaHostname(), this, blockToken);
}
+ protected void copyBlockCrossNamespace(ExtendedBlock sourceBlk,
+ Token<BlockTokenIdentifier> sourceBlockToken, DatanodeInfo
sourceDatanode,
+ ExtendedBlock targetBlk, Token<BlockTokenIdentifier> targetBlockToken,
+ DatanodeInfo targetDatanode) throws IOException {
+ IOStreamPair pair =
+ DFSUtilClient.connectToDN(sourceDatanode,
getConf().getSocketTimeout(), conf, saslClient,
+ socketFactory, getConf().isConnectToDnViaHostname(), this,
sourceBlockToken);
Review Comment:
Plz reuse `connectToDN` and close connection resources. You can refer to
`inferChecksumTypeByReading`.
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java:
##########
@@ -4388,4 +4409,94 @@ boolean isSlownode() {
public BlockPoolManager getBlockPoolManager() {
return blockPoolManager;
}
+
+ public void copyBlockCrossNamespace(ExtendedBlock sourceBlk, ExtendedBlock
targetBlk,
+ DatanodeInfo targetDn) throws IOException {
+ if (!data.isValidBlock(sourceBlk)) {
+ // block does not exist or is under-construction
+ String errStr =
+ "copyBlock:(" + this.getInfoPort() + ") Can't send invalid block " +
sourceBlk + " "
+ + data.getReplicaString(sourceBlk.getBlockPoolId(),
sourceBlk.getBlockId());
+ LOG.info(errStr);
+ throw new IOException(errStr);
+ }
+ long onDiskLength = data.getLength(sourceBlk);
+ if (sourceBlk.getNumBytes() > onDiskLength) {
+ // Shorter on-disk len indicates corruption so report NN the corrupt
block
+ String msg = "copyBlock: Can't replicate block " + sourceBlk + " because
on-disk length "
+ + onDiskLength + " is shorter than provided length " +
sourceBlk.getNumBytes();
+ LOG.info(msg);
+ throw new IOException(msg);
+ }
+ LOG.info(getDatanodeInfo() + " copyBlock: Starting thread to transfer: " +
"block:"
+ + sourceBlk + " from " + this.getDatanodeUuid() + " to " +
targetDn.getDatanodeUuid()
+ + "(" + targetDn + ")");
Review Comment:
> Hi, @LiuGuH , can use log parameterized messages ("{} xxxx", a) instead of
string concatenation (a + xxx).
@LiuGuH Plz fix this.
> Add a new method copyBlockCrossNamespace to DataNode
> ----------------------------------------------------
>
> Key: HDFS-16757
> URL: https://issues.apache.org/jira/browse/HDFS-16757
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: ZanderXu
> Assignee: liuguanghua
> Priority: Minor
> Labels: pull-request-available
>
> Add a new method copyBlockCrossNamespace in DataTransferProtocol at the
> DataNode Side.
> This method will copy a source block from one namespace to a target block
> from a different namespace. If the target DN is the same with the current DN,
> this method will copy the block via HardLink. If the target DN is different
> with the current DN, this method will copy the block via TransferBlock.
> This method will contains some parameters:
> * ExtendedBlock sourceBlock
> * Token<BlockTokenIdentifier> sourceBlockToken
> * ExtendedBlock targetBlock
> * Token<BlockTokenIdentifier> targetBlockToken
> * DatanodeInfo targetDN
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
