[
https://issues.apache.org/jira/browse/RANGER-3961?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Mahesh Hanumant Bandal resolved RANGER-3961.
--------------------------------------------
Fix Version/s: 3.0.0
Resolution: Fixed
This patch is merged into master branch.
Commit id :
[56b241b1472541c280e19219ef65d73413bef2a2|https://github.com/apache/ranger/commit/56b241b1472541c280e19219ef65d73413bef2a2]
> AuditFileSpool logs out all events that were not audited successfully
> ---------------------------------------------------------------------
>
> Key: RANGER-3961
> URL: https://issues.apache.org/jira/browse/RANGER-3961
> Project: Ranger
> Issue Type: Improvement
> Components: Ranger
> Affects Versions: 3.0.0, 2.3.0
> Reporter: Mahesh Hanumant Bandal
> Assignee: Mahesh Hanumant Bandal
> Priority: Major
> Fix For: 3.0.0
>
>
> In AuditFileSpool.java from line 901-905, there is this code:
> {code:java}
> } catch (Throwable t) {
> logger.error("Error while sending logs to consumer. provider="
> + queueProvider.getName() + ", consumer="
> + consumerProvider.getName() + ", log=" + lines, t);
> } {code}
> Here the variable *lines* holds all the events that are in current batch,
> that is 1000 by default (~0.5MB of strings). The batch can be configured even
> higher that makes things worse.
> If there is an issue in audit, a lot of huge strings will be logged out.
> Suggesting to remove *lines* variable and/or add lines.size() to know how
> many events were not sent.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
