You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Gopal V (JIRA)" <ji...@apache.org> on 2015/09/25 22:29:04 UTC
[jira] [Commented] (HIVE-11908) LLAP: Merge branch to hive-2.0
[ https://issues.apache.org/jira/browse/HIVE-11908?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14908628#comment-14908628 ]
Gopal V commented on HIVE-11908:
--------------------------------
Patch applies cleanly onto master & runs queries at 30Tb scale - I've followed the branch throughout, LGTM +1.
As expected, the biggest winners seem to be short BI queries. Something like query55 gets a ~3x boost (150s -> 54s) & query27 gets a ~1.9x boost (250s -> 132s).
At 200Gb scale, the BI wins are more relevant (Query55 summary, for example runs on LLAP under a second easily, but the stats fetch + planner takes ~1800ms).
{code}
INFO : Status: DAG finished successfully in 0.58 seconds
INFO :
INFO : VERTICES TOTAL_TASKS FAILED_ATTEMPTS KILLED_TASKS DURATION_SECONDS CPU_TIME_MILLIS GC_TIME_MILLIS INPUT_RECORDS OUTPUT_RECORDS
INFO : Map 1 1 0 0 0.00 40 0 10,000 31
INFO : Map 2 31 0 0 0.20 45,730 0 18,630,711 4,898
INFO : Map 5 1 0 0 0.00 60 0 48,000 421
INFO : Reducer 3 1 0 0 0.17 170 0 4,898 100
INFO : Reducer 4 1 0 0 0.00 80 0 100 0
{code}
[~sershe]: one zero byte file is missing in the patches needed by the services WebApp.
https://github.com/apache/hive/blob/llap/llap-server/src/main/resources/webapps/llap/.keep
> LLAP: Merge branch to hive-2.0
> ------------------------------
>
> Key: HIVE-11908
> URL: https://issues.apache.org/jira/browse/HIVE-11908
> Project: Hive
> Issue Type: Sub-task
> Affects Versions: 2.0.0
> Reporter: Gopal V
> Assignee: Gopal V
> Priority: Critical
> Labels: TODOC-LLAP
> Attachments: HIVE-11908.patch
>
>
> Merge LLAP branch to hive-2.0.0 (only).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)