You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@phoenix.apache.org by "Istvan Toth (Jira)" <ji...@apache.org> on 2021/01/15 11:59:00 UTC
[jira] [Updated] (PHOENIX-5558) Eliminate the second single data
row scan during read repairs
[ https://issues.apache.org/jira/browse/PHOENIX-5558?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Istvan Toth updated PHOENIX-5558:
---------------------------------
Fix Version/s: 5.1.0
> Eliminate the second single data row scan during read repairs
> -------------------------------------------------------------
>
> Key: PHOENIX-5558
> URL: https://issues.apache.org/jira/browse/PHOENIX-5558
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 5.1.0
> Reporter: Kadir OZDEMIR
> Assignee: Kadir OZDEMIR
> Priority: Major
> Fix For: 5.1.0
>
> Attachments: PHOENIX-5558.master.001.patch
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Initially GlobalIndexChecker used a raw scan to rebuild an index row in order to replay delete markers while rebuilding the row. Although a single index row was rebuilt, the number of mutations that are replayed could be one or more. Because of this, GlobalIndexChecker could not determine if the data table row corresponding to the unverified index row exists or not by just looking at the mutation count returned by UngroupedAggregateRegionObserver. Thus, GlobalIndexChecker had to do another single row scan on the data table to find out if the data table row exist. After PHOENIX-5535, GlobalIndexChecker does not use raw scans anymore. This means that the second single data table row scan during the read repair can be eliminated.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)