Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Hive >> mail # dev >> Re: Review Request 14916: HIVE-5519


+
Hari Sankar Sivarama Subr... 2013-10-25, 19:28
+
Hari Sankar Sivarama Subr... 2013-10-25, 22:56
+
Hari Sankar Sivarama Subr... 2013-10-29, 00:23
Copy link to this message
-
Re: Review Request 14916: HIVE-5519

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/14916/#review27545
-----------------------------------------------------------

http://svn.apache.org/repos/asf/hive/trunk/hcatalog/webhcat/svr/src/main/java/org/apache/hive/hcatalog/templeton/Server.java
<https://reviews.apache.org/r/14916/#comment53463>

    The expected format of starttime should be documented (as well as it's intent)
    
    Also, by reading the code, I'm guessing that the expected use is that when the user wants page 2 of the list, they should supply the last 'starttime' from previous page.  This needs to be documented.
    
    If 'numrecords' == -1, then 'starttime' is not compared with actual job IDs, so the user will get the full list...; that is not what the comment says.
    
    If numrecords > 0, but starttime == null, currRecord is never incremented so the user will get a full list.  Is this intended behavior?
    
    I think there should be some checks for what is valid input and it should throw BadParam when it's not.

http://svn.apache.org/repos/asf/hive/trunk/hcatalog/webhcat/svr/src/main/java/org/apache/hive/hcatalog/templeton/Server.java
<https://reviews.apache.org/r/14916/#comment53464>

    Have you done any profiling wrt how long ListDelegator.run() takes vs fetching detailed info for each job?  If the cluster has 1000s of jobs, is getting a complete list quick or could this also cause a timeout?

http://svn.apache.org/repos/asf/hive/trunk/hcatalog/webhcat/svr/src/main/java/org/apache/hive/hcatalog/templeton/Server.java
<https://reviews.apache.org/r/14916/#comment53466>

    Should this perhaps raise an error for bad input (non-positive number)?
- Eugene Koifman
On Oct. 25, 2013, 7:28 p.m., Hari Sankar Sivarama Subramaniyan wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/14916/
> -----------------------------------------------------------
>
> (Updated Oct. 25, 2013, 7:28 p.m.)
>
>
> Review request for hive, Daniel Dai, Eugene Koifman, and Thejas Nair.
>
>
> Bugs: HIVE-5519
>     https://issues.apache.org/jira/browse/HIVE-5519
>
>
> Repository: hive
>
>
> Description
> -------
>
> Added 2 additional parameters :
> 1. numrecords = number of records the user wants to retrieve based on timestamp ordering
> 2. starttime = start time prefix
>
>
> Diffs
> -----
>
>   http://svn.apache.org/repos/asf/hive/trunk/hcatalog/src/test/e2e/templeton/tests/jobstatus.conf 1535585
>   http://svn.apache.org/repos/asf/hive/trunk/hcatalog/webhcat/svr/src/main/java/org/apache/hive/hcatalog/templeton/Server.java 1535585
>
> Diff: https://reviews.apache.org/r/14916/diff/
>
>
> Testing
> -------
>
> Added e2e test and tested with curl.
>
>
> Thanks,
>
> Hari Sankar Sivarama Subramaniyan
>
>