[ 
https://issues.apache.org/jira/browse/HDFS-9831?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15166336#comment-15166336
 ] 

Xiaoyu Yao commented on HDFS-9831:
----------------------------------

Thanks [~xiaobingo] for working on this. The patch looks good to me. One 
suggestion: can you add some description of the use cases that need to enable 
the WebHDFS retry policy in hdfs-site.xml. For example,

If "true", enable the retry policy of WebHDFS client. This can be useful when 
using WebHDFS to 
 - copy large files between clusters that could timeout or 
 - copy files between HA clusters that could failover during the copy. 


{code}
2834    <property>
2835      <name>dfs.http.client.retry.policy.enabled</name>
2836      <value>false</value>
2837      <description>
2838        If "true", enable the retry policy of WebHDFS client.
2839        If "false", retry policy is turned off.
2840      </description>
2841    </property>
{code}

> Document webhdfs retry configuration keys introduced by HDFS-5219/HDFS-5122 
> ----------------------------------------------------------------------------
>
>                 Key: HDFS-9831
>                 URL: https://issues.apache.org/jira/browse/HDFS-9831
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: documentation, webhdfs
>    Affects Versions: 2.6.0
>            Reporter: Xiaoyu Yao
>            Assignee: Xiaobing Zhou
>         Attachments: HDFS-9831.000.patch
>
>
> This ticket is opened to document the configuration keys introduced by 
> HDFS-5219/HDFS-5122 for WebHdfs Retry.  Both hdfs-default.xml and webhdfs.md 
> should be updated with the usage of these keys.
> {code}
> / WebHDFS retry policy
>   public static final String  DFS_HTTP_CLIENT_RETRY_POLICY_ENABLED_KEY = 
> "dfs.http.client.retry.policy.enabled";
>   public static final boolean DFS_HTTP_CLIENT_RETRY_POLICY_ENABLED_DEFAULT = 
> false;
>   public static final String  DFS_HTTP_CLIENT_RETRY_POLICY_SPEC_KEY = 
> "dfs.http.client.retry.policy.spec";
>   public static final String  DFS_HTTP_CLIENT_RETRY_POLICY_SPEC_DEFAULT = 
> "10000,6,60000,10"; //t1,n1,t2,n2,...
>   public static final String  DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_KEY = 
> "dfs.http.client.failover.max.attempts";
>   public static final int    DFS_HTTP_CLIENT_FAILOVER_MAX_ATTEMPTS_DEFAULT = 
> 15;
>   public static final String  DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_KEY = 
> "dfs.http.client.retry.max.attempts";
>   public static final int    DFS_HTTP_CLIENT_RETRY_MAX_ATTEMPTS_DEFAULT = 10;
>   public static final String  DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_KEY = 
> "dfs.http.client.failover.sleep.base.millis";
>   public static final int    DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_BASE_DEFAULT 
> = 500;
>   public static final String  DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_MAX_KEY = 
> "dfs.http.client.failover.sleep.max.millis";
>   public static final int    DFS_HTTP_CLIENT_FAILOVER_SLEEPTIME_MAX_DEFAULT = 
> 15000;
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to