[
https://issues.apache.org/jira/browse/HDFS-11339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15822191#comment-15822191
]
Arpit Agarwal commented on HDFS-11339:
--------------------------------------
Thanks for this improvement [~hanishakoneru]. A couple of comments:
# Casting has higher precedence than multiply, so I think it will always cast
fileIOSamplingFraction to zero when fileIOSamplingFraction < 1. We need
parenthesis around fileIOSamplingFraction.
{code}
(int) fileIOSamplingFraction * Integer.MAX_VALUE
{code}
# The following check should be == 0.
[System#nanoTime|https://docs.oracle.com/javase/7/docs/api/java/lang/System.html#nanoTime()]
can return negative values.
{code}
if (isEnabled && begin > 0)
{code}
> Support File IO sampling for Datanode IO profiling hooks
> --------------------------------------------------------
>
> Key: HDFS-11339
> URL: https://issues.apache.org/jira/browse/HDFS-11339
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs
> Reporter: Hanisha Koneru
> Assignee: Hanisha Koneru
> Fix For: 3.0.0-alpha2
>
> Attachments: HDFS-11339.000.patch
>
>
> HDFS-10958 introduces instrumentation hooks around DataNode disk IO and
> HDFS-10959 adds support for profiling hooks to expose latency statistics.
> Instead of profiling all the file IO events, we can sample a fraction of the
> events and profile them. The fraction to be sampled can be configurable.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
