[
https://issues.apache.org/jira/browse/SENTRY-1781?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16047502#comment-16047502
]
Alexander Kolbasov commented on SENTRY-1781:
--------------------------------------------
[~spena] Can you clarify why do you want to keep actual path mappings as part
of {{MHiveMetastoreImage}}? I don't see it actually being used anywhere. It
looks like the only thing you really care about is the hmsImageID which is,
essentially, a generation of the full image. Do you use it for anything?
If you don't use it, the row is very small and you shouldn't worry about
multiple entries and don't need the artiicial singleton with
preventConcurrencyCommits. The table becomes very simple then - it can contain
essentially a single column - the generation ID.
What am I missing?
> Define a schema and API to persist an HMS image ID for new HMS snapshots
> ------------------------------------------------------------------------
>
> Key: SENTRY-1781
> URL: https://issues.apache.org/jira/browse/SENTRY-1781
> Project: Sentry
> Issue Type: Sub-task
> Components: Sentry
> Affects Versions: sentry-ha-redesign
> Reporter: Sergio Peña
> Assignee: Sergio Peña
> Fix For: sentry-ha-redesign
>
> Attachments: SENTRY-1781.1-sentry-ha-redesign.patch,
> SENTRY-1781.2-sentry-ha-redesign.patch,
> SENTRY-1781.3-sentry-ha-redesign.patch, SENTRY-1781.4-sentry-ha-redesign.patch
>
>
> A new identifier should be persisted on the table (MAuthzPathsMapping) that
> contains Hive snapshots to identify if a new snapshot was persisted.
> This would be useful for SENTRY-1760
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
