[
https://issues.apache.org/jira/browse/YARN-11054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18039261#comment-18039261
]
ASF GitHub Bot commented on YARN-11054:
---------------------------------------
github-actions[bot] commented on PR #3834:
URL: https://github.com/apache/hadoop/pull/3834#issuecomment-3550016543
We're closing this stale PR because it has been open for 100 days with no
activity. This isn't a judgement on the merit of the PR in any way. It's just a
way of keeping the PR queue manageable.
If you feel like this was a mistake, or you would like to continue working
on it, please feel free to re-open it and ask for a committer to remove the
stale tag and review again.
Thanks all for your contribution.
> Alleviate LocalJobRunnerMetricName Conflicts
> --------------------------------------------
>
> Key: YARN-11054
> URL: https://issues.apache.org/jira/browse/YARN-11054
> Project: Hadoop YARN
> Issue Type: Improvement
> Affects Versions: 3.3.1
> Reporter: Xingjun Hao
> Priority: Minor
> Labels: pull-request-available
> Fix For: 3.3.1
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> In some scenarios, Sqoop will use LocalJobRuner (YarnLocal mode) to run a lot
> of jobs, assuming 2 million jobs have been run, LocalJobRunner MetricName
> generated by nextInt function is in the range of (0, 2147483647),
> Then the probability of conflict is about 2000/2147483647 = 1/1000, which
> means that an average of 1 task will fail for every 1000 jobs run.
> If LocalJobRunner MetricName is generated by nextLong() whose range is (0,
> 9223372036854775807), considering that Long's range is 1 billion times that
> of Int, the probability of a new MetricName conflicts is also reduced by one
> trillionth times.
> the probability of conflict also goes to 1/500000000 from 1/1000. ( under the
> situation that there are about 200million jobs have been run)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
