[
https://issues.apache.org/jira/browse/HDFS-16735?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17583515#comment-17583515
]
ASF GitHub Bot commented on HDFS-16735:
---------------------------------------
Hexiaoqiao commented on code in PR #4780:
URL: https://github.com/apache/hadoop/pull/4780#discussion_r952455223
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/blockmanagement/HeartbeatManager.java:
##########
@@ -71,6 +71,7 @@ class HeartbeatManager implements DatanodeStatistics {
/** Heartbeat monitor thread. */
private final Daemon heartbeatThread = new Daemon(new Monitor());
private final StopWatch heartbeatStopWatch = new StopWatch();
+ private final int removeBatchNum;
Review Comment:
just suggest to define one more readable variable name, such as
`numOfDeadDatanodesRemove`?
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java:
##########
@@ -557,7 +557,8 @@ public class DFSConfigKeys extends CommonConfigurationKeys {
// This value uses the times of heartbeat interval to define the minimum
value for stale interval.
public static final String DFS_NAMENODE_STALE_DATANODE_MINIMUM_INTERVAL_KEY
= "dfs.namenode.stale.datanode.minimum.interval";
public static final int DFS_NAMENODE_STALE_DATANODE_MINIMUM_INTERVAL_DEFAULT
= 3; // i.e. min_interval is 3 * heartbeat_interval = 9s
-
+ public static final String DFS_NAMENODE_REMOVE_BAD_BATCH_NUM =
"dfs.namenode.remove.bad.batch.num";
Review Comment:
`DFS_NAMENODE_REMOVE_DEAD_DATANODE_BATCHNUM_KEY =
"dfs.namenode.remove.dead.datanode.batchnum";` this will be more readable?
> Reduce the number of HeartbeatManager loops
> -------------------------------------------
>
> Key: HDFS-16735
> URL: https://issues.apache.org/jira/browse/HDFS-16735
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Shuyan Zhang
> Assignee: Shuyan Zhang
> Priority: Major
> Labels: pull-request-available
>
> HeartbeatManager only processes one dead datanode (and failed storage) per
> round in heartbeatCheck(), that is to say, if there are ten failed storages,
> all datanode states need to be scanned 10 times, which is unnecessary and a
> waste of resources. This patch makes the number of bad storages processed per
> scan configurable.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
