[jira] [Commented] (HBASE-1422) Refactor to Server Manager

["Hudson (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-1422?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15043525#comment-15043525 ] 

Hudson commented on HBASE-1422:
-------------------------------

SUCCESS: Integrated in HBase-1.3 #420 (See [https://builds.apache.org/job/HBase-1.3/420/])
HBASE-1422 Delayed flush doesn't work causing flush storms; addendum by (stack: rev 7d0c08fbcdb6ab84bc8fb78fdc840c7b29720390)
* hbase-common/src/main/java/org/apache/hadoop/hbase/JitterScheduledThreadPoolExecutorImpl.java


> Refactor to Server Manager
> --------------------------
>
>                 Key: HBASE-1422
>                 URL: https://issues.apache.org/jira/browse/HBASE-1422
>             Project: HBase
>          Issue Type: Sub-task
>    Affects Versions: 0.19.2
>            Reporter: Evgeny Ryabitskiy
>            Assignee: Evgeny Ryabitskiy
>             Fix For: 0.90.0
>
>         Attachments: HBASE-1422.patch, HBASE-1422_v2.patch, HBASE-1422_v3.patch
>
>
> This is refactor to Server Manager class from  HBASE-1017
> I separate it for reasons:
>  * Its better to have several small patchs and apply them iterativly then one great path
>  * I fu..**  tired from synchronising w/ SVN  (this class changes frequently), you can saw 10 patches in HBASE-1017
>  
> We need this refactoing for reasons:
>  *  Server Manager looks like shi**.. bad thing...
>  * is every time harder to make any chnages
>  * it is becoming more ugly every time........
> What changes are done:
> ServerManager has mapping:
>     * serverName 2 serverInfo,
>     * serverAddr 2 serverInfo,
>     * serverName 2 load,
>     * load 2 severName
> 1) serverName 2 load - not necessary if you have serverName 2 serverInfo
> 2) All mappings are encapsulated in ServersInfo class (inner class of ServerManager)
> 3) ServersInfo has operations for adding, updating and removing information of HRS
> + some code in RegionServer is puted in synchronised block... cause it is working with synchronised map...
> Note: this task is to make code much much more clear.. and it's not going to change logic, so no much problem is going appear



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