hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lefty Leverenz (JIRA)" <>
Subject [jira] [Commented] (HIVE-5627) Document 'usehcatalog' parameter on WebHCat calls
Date Sat, 09 Nov 2013 05:00:17 GMT


Lefty Leverenz commented on HIVE-5627:

Missed the update removing "hive" from the list, sorry.

The rest of my comment just points out inconsistencies to fix in the javadocs (internal variable
names should not appear there) and omissions in the Map<String, Object> code that might
or might not need fixing.  Plus a question of whether "enablelog" needs to be documented,
but that's just something I noticed while researching "usehcatalog" so it doesn't belong in
this ticket.  Considering how many inconsistencies I found in these three calls, the other
calls should probably be checked for accuracy too.

Does the information about shipping the Hive tar to the target node only apply to mapreduce/jar
and pig when "usehcatalog" is true, or is it more general information?

> Document 'usehcatalog' parameter on WebHCat calls
> -------------------------------------------------
>                 Key: HIVE-5627
>                 URL:
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Documentation, WebHCat
>    Affects Versions: 0.12.0
>            Reporter: Eugene Koifman
>            Assignee: Lefty Leverenz
>             Fix For: 0.13.0
> The following REST calls in WebHCat:
> 1. mapreduce/jar
> 2. pig
> now support an additional parameter 'usehcatalog'.  This is a mechanism for the caller
to tell WebHCat that the submitted job uses HCat, and thus needs to access the metastore,
which requires additional steps for WebHCat to perform in a secure cluster.  
> The JavaDoc on corresponding methods in  org.apache.hive.hcatalog.templeton.Server describe
this parameter.  
> Additionally, if templeton.hive.archive, templeton.hive.home and templeton.hcat.home
are defined in webhcat-site.xml (documented in webhcat-default.xml) then WebHCat will ship
the Hive tar to the target node where the job actually runs.  This means that Hive doesn't
need to be installed on every node in the Hadoop cluster.  (This part was added in HIVE-5547).
 This is independent of security, but improves manageability.
> This should be added to the sections in
that correspond to these methods.

This message was sent by Atlassian JIRA

View raw message