Pentaho

Not anyone with experience in discovering PDI repositories from a Linux command line tool?

My post is about the problems I'm having when interacting with a Pentaho repository. As I said, I’m unable to run the transformations and/or jobs located in that repository. And you are suggesting to me to run a transformation... Great!

In fact, my problems come earlier than trying to run transformations. As I said, I'm unable to read the directories of a repository. And If I cannot read the directories where the tranformations are stored, how do you expect to invoke tranformations?

Im sorry, I miss read it... Iand assume that you was dealing with some wierd problem with the "repository list" and permissions...

But what sort of problem could it be? As I said, it all works fine when the commands are run on windows based enviroments. The problems only arise on Linux based enviroments. I don't know if the fact that the client and the server are installed on the same machine has to do with it.

Ok, lets slice the problem...

​

Some times "listdir" dont work even when you can connect and execute job/trans.

1 - Can you reach the rep server from linux pdi "https://pentaho.uites.isciii.es" (telnet pentaho.uites.isciii.es 443 / openssl  -s_client...)

2 - Do your java trust the cert from pentaho.uites.isciii.es (cacert)?

3 - can you make a simple (ktr) to connect to the "https://pentaho.uites.isciii.es" and comsume the API?

4 - Can you execute a ktr/kjb on the linux (exported from your windows)?

5 - Can you execute a transformation at the "root" (/) of your repository?

(I know that you said that can't but post it pls)

First of all, I'd like to thank you the time you're investing in trying to help me out.

Regarding the "listdir" command, I'd like to clarify that it only works on Windows environments. On Linux environments it's never worked.

(1) I can reach Pentaho server through both https and telnet protocols:

https://pentaho.uites.isciii.es (browser on Linux and Windows systems)

telnet pentaho.uites.isciii.es 443 (Linux command-line tool)

(2) I don't know how to know if my java trust the cert from my Pentaho server. Could you instruct to me how to find it out?

(3) Could you instruct to me how to do what you're suggesting? I think it has to be done within Pentaho-Spoon but... what step should I use? (I'm really sorry but I'm a newcomer to the Pentaho-Kettle world).

(4) I can confirm that I'm able to run ktr/kjb on Linux environment which had previously been exported from Windows environment. The command I've used is the following:

pan.sh -norep -file <myktrfile>

(5) As far as I know (it may be I'm wrong), Pentaho repositories are sandboxed, that is, you don't know where your repository is located. As a consequence, you just cannot invoke ktr/kjb located at the root of the repository.

Looking forward to reading a reply from you.

Thanks.

2) Its common task having to add certificates to cacerts on https usage, so you have to find where is the cacerts been used by your java install (it might have more then one) and add the certificate to it. To accomplish that, you can use "keytools" to check the cacerts and add the cert to it if needed.

https://knowledge.digicert.com/solution/SO4085.html

https://plone.lucidsolutions.co.nz/linux/java/how-to-add-a-certificate-authority-ca-certificate-to-the-openjdk-cacerts

https://www.sslshopper.com/article-most-common-java-keytool-keystore-commands.html

3) You can use a simple rest step(Or any other http), if your spoon,pan,kitchen are able to reach the server with it... it should be able to connect to your repo.

5) Can be any job/trans on any dir... just need to be one that you are sure where it is... so... somethings in the line:

 ./kitchen.sh -rep=myRep -user=MyUseradmin -pass=MyPass -dir=/home/MyUseradmin -job=myJob_test

In an attempt to simplify the problem I've decided to remove the security in the server. Then, I've changed the content of the $(USER_HOME)/.kettle/repositories.xml so that the <repository_location_url> element takes value of http://pentaho.uites.isciii.es/pentaho. After restarting the Pentaho server and launching the followind command, I get the following error:

$ kitchen.sh -rep:myRepository -user:admin -pass:password -listdir

As I'm able to connect to the pentaho server through http and telnet protocols (http://pentaho.uites.isciii.es using a browser, and typing the command 'telnet pentaho.uites.isciii.es 8080), I don't still know why I'm not able to get the list of directories set in the Pentaho repository.

Any idea/suggestion? Thanks.

Port problem?

You state : "using a browser, and typing the command 'telnet pentaho.uites.isciii.es 8080"

But the connection is trying: "http://pentaho.uites.isciii.es/" (80)

Your repository conf. might be missing the 8080?

All these commands run ok:

• telnet pentaho.uites.isciii.es 80
• telnet pentaho.uites.isciii.es 8080
• http://pentaho.uites.isciii.es

Here it is the content of the $(USER_HOME)/.kettle/repositories.xml file:

<?xml version="1.0" encoding="UTF-8"?>

<repositories>

 <repository>

    <id>PentahoEnterpriseRepository</id>

  <name>myRepository</name>

  <description>PentahoRepository@http://pentaho.uites.isciii.es</description>

  <is_default>true</is_default>

  <repository_location_url>http://pentaho.uites.isciii.es/pentaho</repository_location_url>

  <version_comment_mandatory>N</version_comment_mandatory>

 </repository>

</repositories>

As it can be seen, there is no trace about port configuration whatsoever.

Try the change:

FROM:

<repository_location_url>http://pentaho.uites.isciii.es/pentaho</repository_location_url>

TO:

<repository_location_url>http://pentaho.uites.isciii.es:8080/pentaho</repository_location_url>

Great!! Got it to work!! Really appreciate your help and the time invested.

I don't achieve to understand why the former configuration of the repositories.xml file was valid to discover the repositories created on the server but it wasn't valid to retrieve the directories created on that repo.

Now, with the configuration you suggested, I achieve to retrieve the directories structure on the repository, the list of jobs and/or transformations there stored and, what is more, I'm able to execute these transformations remotely from my Linux command line tool, what is exactly what I wanted.

Thank you very much indeed again!

Which is named in a way to make the version clear UPSers.