phoenix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kadir OZDEMIR (Jira)" <j...@apache.org>
Subject [jira] [Updated] (PHOENIX-5991) IndexRegionObserver should not overwrite mutation timestamps set by clients
Date Tue, 07 Jul 2020 16:26:00 GMT

     [ https://issues.apache.org/jira/browse/PHOENIX-5991?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Kadir OZDEMIR updated PHOENIX-5991:
-----------------------------------
    Attachment:     (was: -PHOENIX-5991.4.x.001.patch)

> IndexRegionObserver should not overwrite mutation timestamps set by clients
> ---------------------------------------------------------------------------
>
>                 Key: PHOENIX-5991
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-5991
>             Project: Phoenix
>          Issue Type: Improvement
>            Reporter: Kadir OZDEMIR
>            Assignee: Kadir OZDEMIR
>            Priority: Major
>             Fix For: 4.x
>
>         Attachments: PHOENIX-5991.4.x.001.patch
>
>          Time Spent: 40m
>  Remaining Estimate: 0h
>
> Phoenix client can set mutation timestamps for connections with scn. For example, it
sets the timestamp for delete mutations. IndexRegionObserver does not check if the timestamp
is set for mutations and overwrites them with its current time. While Phoenix does not support
setting arbitrary timestamps on the client side for all operations, setting a timestamp for
delete mutations is harmless even for tables with indexes. Instead of blindly overwriting
mutations timestamps, IndexRegionObserver can leave the determination of which timestamp values
are allowed for which operations and sets the timestamp for mutations only when they are not
set by Phoenix client.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Mime
View raw message