[
https://issues.apache.org/jira/browse/HIVE-3707?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kevin Wilfong updated HIVE-3707:
--------------------------------
Attachment: HIVE-3707.1.patch.txt
> Round map/reduce progress down when it is in the range [99.5, 100)
> ------------------------------------------------------------------
>
> Key: HIVE-3707
> URL: https://issues.apache.org/jira/browse/HIVE-3707
> Project: Hive
> Issue Type: Improvement
> Components: Logging, Query Processor
> Affects Versions: 0.10.0
> Reporter: Kevin Wilfong
> Assignee: Kevin Wilfong
> Priority: Minor
> Attachments: HIVE-3707.1.patch.txt
>
>
> In HadoopJobExecHelper the mapProgress and reduceProgress are the value of
> these counters taken from the running job rounded to an integer percentage.
> This means that e.g. if the mappers are 99.5% done this is stored as 100%.
> One of the most common questions I see from new users is, the map and reduce
> both report being 100% done, why is the query still running?
> By rounding down the value in this interval so it's only 100% when it's
> really 100% we could avoid that confusion.
> Also, the way it appears the QueryPlan and MapRedTask determine if the
> map/reduce phases are done is by checking if this value == 100. I couldn't
> find anywhere where they're used for anything significant, but they're
> reporting early completion.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
