[
https://issues.apache.org/jira/browse/HDFS-17471?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17841880#comment-17841880
]
ASF GitHub Bot commented on HDFS-17471:
---------------------------------------
Hexiaoqiao commented on code in PR #6742:
URL: https://github.com/apache/hadoop/pull/6742#discussion_r1582649920
##########
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/ProfilingFileIoEvents.java:
##########
@@ -80,7 +80,7 @@ public void afterMetadataOp(@Nullable FsVolumeSpi volume,
public long beforeFileIo(@Nullable FsVolumeSpi volume,
FileIoProvider.OPERATION op, long len) {
- if (isEnabled && ThreadLocalRandom.current().nextInt() < sampleRangeMax) {
+ if (isEnabled && ThreadLocalRandom.current().nextInt(Integer.MAX_VALUE) <
sampleRangeMax) {
Review Comment:
Great catch! some other same logic should be corrected too, such as
org.apache.hadoop.hdfs.client.impl.metrics.BlockReaderIoProvider#read, please
also check them together.
> Correct the percentage of file I/O events.
> ------------------------------------------
>
> Key: HDFS-17471
> URL: https://issues.apache.org/jira/browse/HDFS-17471
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs
> Reporter: fuchaohong
> Priority: Major
> Labels: pull-request-available
>
> The percentage of file I/O events to be profiled for DataNode disk statistics
> is inaccurate.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
