[jira] [Comment Edited] (RANGER-838) Tag-sync should be resilient to Ranger Admin availability

["Madhan Neethiraj (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/RANGER-838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15227111#comment-15227111 ] 

Madhan Neethiraj edited comment on RANGER-838 at 4/5/16 8:47 PM:
-----------------------------------------------------------------

+1 for the patch.

Patch committed to master: http://git-wip-us.apache.org/repos/asf/incubator-ranger/commit/93cb024d


was (Author: madhan.neethiraj):
+1 for the patch

> Tag-sync should be resilient to Ranger Admin availability
> ---------------------------------------------------------
>
>                 Key: RANGER-838
>                 URL: https://issues.apache.org/jira/browse/RANGER-838
>             Project: Ranger
>          Issue Type: Bug
>          Components: Ranger
>    Affects Versions: 0.6.0
>            Reporter: Madhan Neethiraj
>            Assignee: Abhay Kulkarni
>             Fix For: 0.6.0
>
>         Attachments: 0001-RANGER-838-Tag-sync-should-be-resilient-to-Ranger-Ad.patch
>
>
> If Ranger Admin is not reachable during startup of tag-sync daemon, the daemon exits currently. Until the daemon is started again manually, Ranger will not get tag information updates from Apache Atlas (or other configured tag sources).
> Ranger tag-sync should be able to handle failures in connecting to external systems it interacts with - like Ranger Admin/Kafka/Atlas, without aborting. For example:
>  - if Ranger Admin is not reachable, Kafka notification processing should be paused and should be resumed only after Ranger Admin is reachable
>  - on failures in receiving notification from Kafka, appropriate wait/retry should be in place to be able to continue once Kafka notifications arrive
>  - similarly for failures in reaching Apache Atlas



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)