> On July 13, 2015, 10:48 a.m., Shwetha GS wrote:
> > For hive hook, there is a config:
> > <property>
> >   <name>atlas.rest.address</name>
> >   <value>http://localhost:21000/</value>
> > </property>
> > 
> > 
> > Can we use the same config to check if its ssl or not, get the port and 
> > also bind to an address?
> 
> Jonathan Maron wrote:
>     We might be able to do that, but from a management interface it may not 
> be ideal and would probably run contrary to the way in which users expect to 
> manage the assignment of ports (see yarn and hdfs as examples).  I think it 
> is preferable to have persisted values for each transport type since that is 
> a familiar and well-understood mechanism for configuration in a Hadoop 
> cluster.
> 
> Shwetha GS wrote:
>     Even hdfs and yarn use the full endpoint configuration to bind the port
>      <property>
>         <description>The address of the applications manager interface in the 
> RM.</description>
>         <name>yarn.resourcemanager.address</name>
>         <value>${yarn.resourcemanager.hostname}:8032</value>
>       </property>
>       <property>
>         <description>The hostname of the RM.</description>
>         <name>yarn.resourcemanager.hostname</name>
>         <value>0.0.0.0</value>
>       </property>
> 
> Jonathan Maron wrote:
>     I was talking specifically about the pattern of enabling SSL via a flag 
> (which we do), and defining an associated port, much like HDFS, for example:
>     
>         <property>
>           <name>dfs.http.policy</name>
>           <value>HTTPS_ONLY</value>
>         </property>
>     
>         <property>
>           <name>dfs.https.port</name>
>           <value>50470</value>
>         </property>
>     
>     Given the time differences and its effect on reviews (this simple change 
> has now been in review for 6 days) I think that:
>     
>     1)  This is probably not a valid issue to prolong this review
>     2)  If users complain we can certainly address a change of the approach 
> with a subsequent JIRA.
>     
>     I will therefore proceed with a commit.

I guess the reason it has this config is because it hosts server on both http 
and https at the same time. For our case, we enabled either http or https, it 
makes sense to read from url. Will create another bug


- Shwetha


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/36359/#review91458
-----------------------------------------------------------


On July 9, 2015, 6:06 p.m., Jonathan Maron wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/36359/
> -----------------------------------------------------------
> 
> (Updated July 9, 2015, 6:06 p.m.)
> 
> 
> Review request for atlas, Arpit Gupta and Seetharam Venkatesh.
> 
> 
> Bugs: ATLAS-36
>     https://issues.apache.org/jira/browse/ATLAS-36
> 
> 
> Repository: atlas
> 
> 
> Description
> -------
> 
> Distinct HTTP and HTTPS persisted port properties will make the configuration 
> clearer and allow for a better experience from management interfaces such as 
> Ambari.
> 
> 
> Diffs
> -----
> 
>   src/conf/application.properties 6c4c7d2 
>   webapp/src/main/java/org/apache/atlas/Main.java 1dd17b4 
>   webapp/src/test/java/org/apache/atlas/MainIT.java PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/36359/diff/
> 
> 
> Testing
> -------
> 
> Installed a local server and observed the configuration properties working.
> Added a unit test for port selection logic (MainIT).
> 
> 
> Thanks,
> 
> Jonathan Maron
> 
>

Reply via email to