[jira] [Resolved] (TEPHRA-230) Improve the logic used to find recorded regions during invalid list pruning

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

     [ https://issues.apache.org/jira/browse/TEPHRA-230?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Poorna Chandra resolved TEPHRA-230.
-----------------------------------
    Resolution: Fixed

> Improve the logic used to find recorded regions during invalid list pruning
> ---------------------------------------------------------------------------
>
>                 Key: TEPHRA-230
>                 URL: https://issues.apache.org/jira/browse/TEPHRA-230
>             Project: Tephra
>          Issue Type: Bug
>    Affects Versions: 0.10.0-incubating
>            Reporter: Poorna Chandra
>            Assignee: Poorna Chandra
>             Fix For: 0.12.0-incubating
>
>
> During an invalid list pruning run, the plugin records the regions present on the HBase cluster at that time. It also records the count of regions as a checksum. 
> While reading the list of recorded regions, if the checksum does not match the number of regions read at time {{t}}, then the plugin will move on to regions at time {{t - 1}}. However, the {{t-1}} computation in inefficient today, and leads to microsecond decrements. It can be improved to use the time at which the current set of regions were recorded.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)