[ 
https://issues.apache.org/jira/browse/SENTRY-2396?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16609235#comment-16609235
 ] 

kalyan kumar kalvagadda edited comment on SENTRY-2396 at 9/10/18 1:54 PM:
--------------------------------------------------------------------------

[~LinaAtAustin] Conceptually SENTRY-2300 will address the 0.5.sec delay in 
fetching the notifications but this approach significantly changes the current 
way of fetching changes to HMS data and the changes are not trivial.  We need 
to look very closely to understand all the effects of doing so.

 

[~akolb] I have plans to address the delay in the Follower node in another Jira 
[SENTRY-2397]. Idea is to increase the frequency of the polls in the Follower 
when it's not in leader. It depends on the HmsFollowerScheduler, introduced in 
this patch.


was (Author: kkalyan):
[~LinaAtAustin] Conceptually SENTRY-2300 will address the 0.5.sec delay in 
fetching the notifications but this approach significantly changes the current 
way of fetching changes to HMS data and the changes are not trivial.  We need 
to look very closely to understand all the effects of doing so.

 

[~akolb] I have plans to address the delay in the Follower node in another Jira 
\[SENTRY-2397\]. Idea is to increase the frequency of the polls in the Follower 
when it's not in leader.

> Minimize the time sentry blocks HMS threads
> -------------------------------------------
>
>                 Key: SENTRY-2396
>                 URL: https://issues.apache.org/jira/browse/SENTRY-2396
>             Project: Sentry
>          Issue Type: Sub-task
>          Components: Sentry
>    Affects Versions: 2.1.0
>            Reporter: kalyan kumar kalvagadda
>            Assignee: kalyan kumar kalvagadda
>            Priority: Major
>         Attachments: SENTRY-2396.001.patch, SENTRY-2396.002.patch
>
>
> Currently HMS is blocked for max of one sec by sentry for below operations
>  # create database
>  # drop database
>  # create table
>  # drop table
>  # alter table
> Reason for this is to make sure that sentry fetched  the notification event 
> from HMS and processed.
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to