[
https://issues.apache.org/jira/browse/HDFS-17044?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17733715#comment-17733715
]
ASF GitHub Bot commented on HDFS-17044:
---------------------------------------
Hexiaoqiao commented on code in PR #5735:
URL: https://github.com/apache/hadoop/pull/5735#discussion_r1232990215
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockManager.java:
##########
@@ -3295,8 +3295,11 @@ private BlockInfo processReportedBlock(
BlockInfo storedBlock = getStoredBlock(block);
if(storedBlock == null) {
// If blocksMap does not contain reported block id,
- // the replica should be removed from the data-node.
- toInvalidate.add(new Block(block));
+ // the replica should be removed from the data-node,
Review Comment:
The replica should be removed from Datanode, and set NumBytes to
BlockCommand.No_ACK to avoid useless report to NameNode from Datanode when
complete to process it.
> Process reported block toInvalidate logic should set the block size to NO_ACK
> -----------------------------------------------------------------------------
>
> Key: HDFS-17044
> URL: https://issues.apache.org/jira/browse/HDFS-17044
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Haiyang Hu
> Assignee: Haiyang Hu
> Priority: Major
> Labels: pull-request-available
>
> When NameNode processes DataNode increment or full block report, if block is
> not in the blocks map, it will be added to invalidate and the replica should
> be removed from the data-node, and the block size should be to set NO_ACK,
> should be reduce some useless DataNode block reports.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
