[
https://issues.apache.org/jira/browse/HDFS-12502?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16217740#comment-16217740
]
Daryn Sharp commented on HDFS-12502:
------------------------------------
I'd prefer to see it reverted while we figure out how to add it differently.
We've already had to internally revert this change.
I have a general concern about the new metrics creep. Nothing is free... I
don't want Heisenberg to become an unwelcome roommate. Out of curiosity, can
you share general details of the incident that motivates you to add this
metric? How much of an impact did you see from listing a large dir? Are you
using the fair call queue? I'm far more worried about a 10k+ create/sec flood
than a 100k+ listStatus/sec flood regardless of the number of items in the dir.
> nntop should support a category based on FilesInGetListingOps
> -------------------------------------------------------------
>
> Key: HDFS-12502
> URL: https://issues.apache.org/jira/browse/HDFS-12502
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: metrics
> Reporter: Zhe Zhang
> Assignee: Zhe Zhang
> Fix For: 2.9.0, 2.8.3, 2.7.5, 3.0.0, 3.1.0
>
> Attachments: HDFS-12502.00.patch, HDFS-12502.01.patch,
> HDFS-12502.02.patch, HDFS-12502.03.patch, HDFS-12502.04.patch
>
>
> Large listing ops can oftentimes be the main contributor to NameNode
> slowness. The aggregate cost of listing ops is proportional to the
> {{FilesInGetListingOps}} rather than the number of listing ops. Therefore
> it'd be very useful for nntop to support this category.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
