[ 
https://issues.apache.org/jira/browse/HDFS-5776?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13883928#comment-13883928
 ] 

Jing Zhao commented on HDFS-5776:
---------------------------------

bq.  it's more flexible if we provide instance level disable/enable APIs, so we 
can archive to use the hbase shell script to control the switch per dfs client 
instance, that'll be cooler
I still have some concern about the current implementation: 
1) we do not check threadpool in enableHedgedReads. This makes it possible that 
isHedgedReadsEnabled() returns true while hedged read is actually not enabled.
2) DFSClient#setThreadsNumForHedgedReads allows users to keep changing the size 
of the thread pool.
To provide instance level disable/enable APIs, I think maybe we can do the 
following:
1) Read the thread pool size configuration only when initializing the thread 
pool, and the size should be >0 and cannot be changed.
2) Add an "Allow-Hedged-Reads" configuration. Each DFSClient instance reads 
this configuration, and if it is true, checks and initializes the thread pool 
if necessary. Users can turn on/off the switch using the enable/disable 
methods. In the enable method, we check and initialize the thread pool if 
necessary.

What do you think [~xieliang007]?

> Support 'hedged' reads in DFSClient
> -----------------------------------
>
>                 Key: HDFS-5776
>                 URL: https://issues.apache.org/jira/browse/HDFS-5776
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: hdfs-client
>    Affects Versions: 3.0.0
>            Reporter: Liang Xie
>            Assignee: Liang Xie
>         Attachments: HDFS-5776-v10.txt, HDFS-5776-v2.txt, HDFS-5776-v3.txt, 
> HDFS-5776-v4.txt, HDFS-5776-v5.txt, HDFS-5776-v6.txt, HDFS-5776-v7.txt, 
> HDFS-5776-v8.txt, HDFS-5776-v9.txt, HDFS-5776.txt
>
>
> This is a placeholder of hdfs related stuff backport from 
> https://issues.apache.org/jira/browse/HBASE-7509
> The quorum read ability should be helpful especially to optimize read outliers
> we can utilize "dfs.dfsclient.quorum.read.threshold.millis" & 
> "dfs.dfsclient.quorum.read.threadpool.size" to enable/disable the hedged read 
> ability from client side(e.g. HBase), and by using DFSQuorumReadMetrics, we 
> could export the interested metric valus into client system(e.g. HBase's 
> regionserver metric).
> The core logic is in pread code path, we decide to goto the original 
> fetchBlockByteRange or the new introduced fetchBlockByteRangeSpeculative per 
> the above config items.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Reply via email to