[
https://issues.apache.org/jira/browse/SENTRY-2114?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16310384#comment-16310384
]
kalyan kumar kalvagadda commented on SENTRY-2114:
-------------------------------------------------
[~lina.li] Even the current code doesn't handle this scenario properly. I was
planning to create another jira to fix the functionality when
1. HDFS is disabled
2. HDFS is enabled after some time.
> Fix behavior of HMSFollower when HDFS sync is not enabled
> ---------------------------------------------------------
>
> Key: SENTRY-2114
> URL: https://issues.apache.org/jira/browse/SENTRY-2114
> Project: Sentry
> Issue Type: Bug
> Components: Sentry
> Affects Versions: 2.1.0
> Reporter: kalyan kumar kalvagadda
> Assignee: kalyan kumar kalvagadda
>
> Currently behavior of HMSFollower when HDFS sync disabled is as below.
> # Take full snapshot of HMS
> # Take the Notification Id from the snapshot created.
> # Persist the notification Id to SENTRY_HMS_NOTIFICATION_ID table
> # Fetches notifications and process new notifications after that point.
> *This doesn't make sense. Ideally, I think it should work as below.*
> # Initially, fetch all the notifications with even-id > 0 and process them.
> # Subsequently fetch and process new notifications.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
