ambari-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mithun Mathew <mithm...@gmail.com>
Subject Re: Step 4 in Adding a custom defined service is getting stuck at Customize Services screen
Date Tue, 19 Apr 2016 07:13:52 GMT
The app.js file gets deployed in /usr/lib/ambari-server/web/javascripts
You can make changes to this file and test it right away - refreshing the
browser (hard-refresh) would reload app.js.

If you wish to edit the actual source code, app.js is a minified version of
ambari-web <https://github.com/apache/ambari/tree/trunk/ambari-web>. You
may look into Frontend development section on Ambari Developer Wiki
<https://cwiki.apache.org/confluence/display/AMBARI/Quick+Start+Guide> for
getting started on this.


On Mon, Apr 18, 2016 at 11:36 PM, Souvik Sarkhel <souvik.sarkhel@gmail.com>
wrote:

> Hi Mithun,
>
> I guess I found out what is causing the error. In app.js line 159865
>
>
>
>
>
>
>
>
> *case 'dfs.data.dir':      case 'dfs.datanode.data.dir':        temp =
> slaveComponentHostsInDB.findProperty('componentName', 'DATANODE');
> temp.hosts.forEach(function (host) {
> setOfHostNames.push(host.hostName);        }, this);*
> if a config file contains *dfs.data.dir *then this piece of code looks
> for slaves of the component named *DATANODE*.
> But in my stack there is no component named *DATANODE *instead its named '
> *HDFS_DATANODE*' hence its giving this error
>
> TypeError: temp is undefined in
>
> temp.hosts.forEach(function (host)
>
> How to solve this ?
>
>
> Where can i find the app.js file so that I can change the value of
> datanode to hdfs_datanode
>
>
>
> On Tue, Apr 19, 2016 at 11:30 AM, Mithun Mathew <mithmatt@apache.org>
> wrote:
>
>> To my understanding configuration file names should be unique across
>> services installed in the cluster.
>> To confirm this, you may open the developer console and see what the
>> error is. If it is stuck on the loading icon, it is highly likely that
>> there is an error thrown by JS, the result of which configs were not loaded.
>>
>> On Mon, Apr 18, 2016 at 10:43 PM, Souvik Sarkhel <
>> souvik.sarkhel@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> I have created a custom service named *HDFSYARN* which installs *Hadoop*
>>> in all the nodes and starts namenode, datanode and resource manager in yarn
>>> mode. I want the user to be able to modify the following .xml files:
>>>
>>>
>>>
>>> *capacity-scheduler.xmlcore-site.xmlmapred-site.xml*
>>>
>>> *yarn-site.xml*
>>> *hdfs-site.xml*
>>>
>>>  I have placed the followed the following folder structure
>>> *metainfo.xml*
>>> |_ *configuration*
>>>
>>> *capacity-scheduler.xml*
>>> *      core-site.xml*
>>> *      mapred-site.xml*
>>> *      yarn-site.xml*
>>> *      hdfs-site.xml*
>>>
>>> *|_ package*
>>>
>>> *        |_ scripts*
>>>
>>> *                 master.py*
>>>
>>> *                  slave.py*
>>>
>>> and my *metainfo.xml* file looks like this
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> *<?xml version="1.0"?><metainfo>  <schemaVersion>2.0</schemaVersion>
>>> <services>    <service>      <name>HDFSYARN</name>  
   <displayName>HDFS
>>> YARN</displayName>      <comment>HDFS is a Java-based file system
that
>>> provides scalable and reliable data storage, it is designed to span large
>>> clusters of commodity servers</comment>      <version>2.6.0</version>
>>> <components>        <component>
>>> <name>HDFS_NAMENODE</name>          <displayName>HDFS
>>> NameNode</displayName>          <category>MASTER</category>
>>> <cardinality>1</cardinality>
>>> <timelineAppid>HDFSYARN</timelineAppid>
>>> <dependencies>                    <dependency>
>>> <name>TOMCAT/TOMCAT_SLAVE</name>
>>> <scope>cluster</scope>
>>> <auto-deploy>
>>> <enabled>true</enabled>
>>> </auto-deploy>                </dependency>
>>> </dependencies>          <commandScript>
>>> <script>scripts/master.py</script>
>>> <scriptType>PYTHON</scriptType>            <timeout>1200</timeout>
>>> </commandScript>        </component>        <component>
>>> <name>HDFS_DATANODE</name>          <displayName>HDFS
>>> DataNode</displayName>          <cardinality>0+</cardinality>
>>> <category>SLAVE</category>
>>> <timelineAppid>HDFSYARN</timelineAppid>          <commandScript>
>>> <script>scripts/slave.py</script>
>>> <scriptType>PYTHON</scriptType>            <timeout>1200</timeout>
>>> </commandScript>        </component>   </components>
>>> <osSpecifics>        <osSpecific>
>>> <osFamily>any</osFamily>          <!-- note: use osType rather
than
>>> osFamily for Ambari 1.5.0 and 1.5.1 -->          <packages>
>>> <package>              <name>hadoop-2.6.0</name>
>>> </package>          </packages>        </osSpecific>
>>> </osSpecifics>      <requiredServices>
>>> <service>TOMCAT</service>      </requiredServices>
>>> <configuration-dependencies>
>>> <config-type>core-site</config-type>
>>> <config-type>hdfs-site</config-type>
>>> <config-type>mapred-site</config-type>
>>> <config-type>capacity-scheduler</config-type>
>>> <config-type>yarn-site</config-type>      </configuration-dependencies>
>>> </service>  </services></metainfo>*
>>>
>>> But the moment I place *hdfs-site.xml* and *yarn-site.xml* in
>>> configuration folder and try to add the service it gets stuck at Customize
>>> Services window
>>> [image: Inline image 1]
>>> and when my configuration folder doesn't contains those two files
>>> everything works properly.
>>> Is it because *HDP* stack also has *HDFS* and *YARN* services and
>>> somehow Ambari is still fetching some dependencies from those services
>>> instead of the custom defined service.?
>>> Thanking you in advance
>>>
>>> --
>>> Souvik Sarkhel
>>>
>>
>>
>
>
> --
> Souvik Sarkhel
>



-- 
*Mithun Mathew* (Matt)

   - www.linkedin.com/in/mithunmatt/

Mime
View raw message