[
https://issues.apache.org/jira/browse/HDFS-17060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17736982#comment-17736982
]
ASF GitHub Bot commented on HDFS-17060:
---------------------------------------
xinglin commented on code in PR #5778:
URL: https://github.com/apache/hadoop/pull/5778#discussion_r1241497840
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/BlockPlacementPolicyDefault.java:
##########
@@ -1236,11 +1245,17 @@ public DatanodeStorageInfo chooseReplicaToDelete(
minSpace = free;
minSpaceStorage = storage;
}
+ if (considerLoad2chooseReplicaDeleting && nodeLoad < minLoad) {
+ minLoad = nodeLoad;
+ minLoadStorage = storage;
+ }
}
final DatanodeStorageInfo storage;
if (oldestHeartbeatStorage != null) {
Review Comment:
this `if else` code section seems problematic: After exiting from the
previous for loop, minSpaceStorage should not be null because minSpace is
initialized to Long.MAX_VALUE. So, `storage = minSpaceStorage;` will always be
executed: it seems we are always picking a replica based on remaining space,
rather than the other two metrics.
> BlockPlacementPolicyDefault#chooseReplicaToDelete should consider datanode
> load
> -------------------------------------------------------------------------------
>
> Key: HDFS-17060
> URL: https://issues.apache.org/jira/browse/HDFS-17060
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: farmmamba
> Assignee: farmmamba
> Priority: Minor
> Labels: pull-request-available
>
> When choose extra replicas for deleting, we should consider datanode load as
> well.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
