[
https://issues.apache.org/jira/browse/RANGER-4170?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17707814#comment-17707814
]
Xiaomin Zhang commented on RANGER-4170:
---------------------------------------
Thanks [~madhan] We are observing this issue in Cloudera specific ranger
release 2.1 but with all the mentioned fixes backported.
> keep policy-cache snapshot for each update
> ------------------------------------------
>
> Key: RANGER-4170
> URL: https://issues.apache.org/jira/browse/RANGER-4170
> Project: Ranger
> Issue Type: Improvement
> Components: plugins
> Reporter: Xiaomin Zhang
> Priority: Major
>
> We observed strange issue that plugin is not able to find the correct policy
> hence failing the access request. This is intermittent and not reproduced at
> will. This issue seems gone after restarting the component or RangerAdmin or
> make some slight changes to the affected policy, or just wait for some while
> without doing any thing.
> Since policy-cache snapshot at the issue time is critical to debug this
> issue, but there is a major gap here to keep the snapshot. We need some
> improvement in component plugin to keep every policy-cache snapshot when it's
> updated, so that we can trace back to find the policy-cache at the point of
> time when issue recurs.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
