Puppet, Augeas and sudoers

I'm developing some puppet modules in these days and I'm working on a Moodle module.

The interesting thing is that Moodle has a nice CLI interface that is possible to use to make the installation unattended and not iterative so is possible to create a puppet module that manages the whole installation by itself.
The module is still in the works, but I'm trying to make it as modular as possible since I could need to have the MySQL DB in a different host and so I prefer to set up the module to delegate to MySQL the task of building the DB.
This is something that many modules around to not take into consideration: not all environments have the DB in the same host as the application you are installing.  In my case is particularly true in my case since I work in a cloud environment and I need the machine to be automatically deployed in CloudStack letting the Moodle machine to export the DB need to the MySQL central machine.

To solve this problem I'm using the exported resources and I can say that it works pretty well as I posted some times ago.

With this approach the Moodle module just takes care of the Moodle stuff and demand (issues a export resource and MySQL will catch it and create the DB) to other modules the task of managing their resources.

Thankfully puppetlabs is working hard to provide a set of good basic modules to manage thinks like MySQL (had to use the GIT version since there is a bug that has been fixed there)
One of the thingsI had a really awful time to solve is the use of Augeas in puppet to configure the /etc/sudoers file.
The problem is that Moodle CLI wants to be executed as the Apache user, but the apache user do not have TTY and is not allower to work without a password so googling around (I'm not a sysop, more a jack-of-all-trades and thus have to google a lot in my day :-) ) it turned out I had to have sudo and write /etc/sudoers to enable apache to work.
It seems easy ... (at least if you use visudo), but using puppet ...
Looking around I felt lucky augeas can write on /etc/sudoers ... but how?  No good example around so after many trials and error I decided to make this post to help others not loosing all the time I did:

augeas { "sudoapache":
    context => "/files/etc/sudoers",
    changes => [
      "set spec[user = '$moodle::params::web_user']/user \"$moodle::params::web_user\"",
      "set spec[user = '$moodle::params::web_user']/host_group/host \"ALL\"",
      "set spec[user = '$moodle::params::web_user']/host_group/command \"ALL\"",
      "set spec[user = '$moodle::params::web_user']/host_group/command/runas_user \"ALL\"",
      "set spec[user = '$moodle::params::web_user']/host_group/command/tag \"NOPASSWD\"",
      "set Defaults[type=':$moodle::params::web_user']/type :$moodle::params::web_user",
      "set Defaults[type=':$moodle::params::web_user']/requiretty/negate \"\""],
  }

This will produce the two rows in the file that allowed the CLI to make its own work.

apache ALL = (ALL) NOPASSWD : ALL
Defaults:apache !requiretty

Good puppetizing!

VUVOX a timeline storytelling

This is a really powerfull solution for alternative form of presentations.

The timeline approach transmits a narrative flow to the presentation and the power of hyperlinks to show videos, images, audio is a big addition.

This is one complementary alternative to Prezi.

You give the user the power to control the flow and the hyperlink is something tha Prezi lacks. These two tools offer a new way to present content in a rich and engaging style.

I can advise it.

Your data in the social network

This video is really funny!!

Is a good example of how our personal information is available to public access.

This true especially for young users of the internet, too many times people publish things that could represent a problem for their current or future public image or a threath to the security of web accounts.

TEIID 8.1 tutorial - Part 5

We prepared a VDB in the previous part of the TEIID tutorial.

Now we want to test it on a TEIID server.

To do so we need a fresh and clean TEIID installation.

We download a JBoss AS version 7.1.1 and unzip it into a folder (I'm runniong the tutorial on windows, but you can do the same on linux).

I usually check that JBoss works out of the box before installing anything.

When you have checked the JBoss AS installation you are ready to install TEIID, that is pretty easy stuff to do as you have to download the TEIID distribution and just unpack it into the JBoss installation; there is the same folder structure so everything will go to its place.

For the demo we'll going to run JBoss in standalone mode, this means that we'll look at the files into the standalone folder.

Since there is a standalone-teiid.xml file I usually create a start script to set the config file to be that one, but is a habit of mine, you can rename it standalone.xml and work on that (just keep the original standalone somewhere should you need it).

Done that just start JBoss again and this time we'll see that also TEIID modules gets loaded and it is ready and running; obviously there is no DB right now.

Shut down everything and make some few configurations (all these stuffs can be done in different ways, as a tutorial I prefer to show what gets written into the XML file so that is known what web consolle does behind the scene).

Since we are going to connect to MySQL we need to prepare the module for MySQL with the driver so that JBoss and TEIID can make a connection to MySQL create a datasource to our DB with the correct jndi name we used in the VDB.

The TEIID installation comes with a usefull set of examples/templates for the modules so we just have to copy the MySQL one in the modules folder, copy the jdbc driver in the folder, configure the modules with the correct name of the lib and the driver is ready.

We than need to add the driver information within the standalone.xml file.

Done that we can add the driver information in the standalone.xml file and set all the parameters to be able to correctly connect to the DB.

When you have finished the configuration start JBoss and go to the web concolle and check the functioning of the datasource testing the connection with the DB.



Once we have finished this part we can deploy the VDB.

Project ReteLIM - thanks to open source

I'm working on a very interesting project called ReteLIM.

Is a project to help hospitalized children to stay in contact with their class and to enable them to work together with their classmates.

I'm working side by side with teachers finding the best technical solution to apply to the task.

The project aims to define which hardware and software use and how to help teachers apply technology to the learning experience.

The model start with hospitalized children constraint, but can be applyed to any "remote" situation.

Our idea is to implement a "virtual classroom" using integration of open source tools to enable a complete learning experience helping a standardization of approach.

We decided to use a cloud hosting to serve the software.

In the cloud the teachers will find a Ubuntu desktop that they will be able to access from anywhere thanks to NX web access.

This desktop is loaded with all sort of program dedicated to education (it could be an Edubuntu distro, but for now we preferred to go with the standard one).

This desktop will also mount an NFS share from a server.

The server hosts an integrated environment composed by Joomla!! + Moodle + Mahara + OpenMeetings (single sign on thanks to Joomdle).

The nfs share will be seen by Moodle as a repository so all the content, produced and saved into a particular folder of the desktop, will be seen and available to Moodle without any trouble for the teacher.

Also OpenMeetings will be available as a Moodle activity so that it will be possible to use it to work with the hospitalized child and the support teacher (the teacher will be equipped with a tablet PC) and the class where a LIM (that is the italian acronymn for an interactive whiteboard) can be used: either the one done with the WII or commercial ones that public administration gave to the school.

The project is fantastic and the energy of the group is vibrant, this is why I choose IT as working life: it can really help making the difference for the lives of many, seeing the happyness in the eyes of the children that can feel to be still part of a normal routine is the greatest satisfaction.

I just feel that is important for me to thanks all the people that work in the open source ecosystem: without their work this project would not be possible and affordable.

Thanks guys.

Evolution of TEIID 8.1 tutorial

I've decided to expand this tutorial to add integration with other pieces of software I'm using in the Open-DAI project I'm working on.

Just to remember what I've already published in the tutorial right now:
1) Create a relational mapping
2) Create a XML document from an XSD schema
3) Map the relational DB to the XML document
4) Create the VDB

Planned coming next:
5) Install TEIID in JBoss AS
6) Deploy the VDB in TEIID server
7) Connect to the VDB with SquirrelSQL and get the XML document with a query
8) Create a web service that will publish the relational tables
9) Deploy the new VDB and the Web Service and test it with soapUI

The tutorial will follow these additional steps to demonstrate the integration of a data service generated with TEIID with other platforms:

10) Publish the TEIID web service with WSO2 API manager
11) test the published API with soapUI
12) publish the relational tables as REST service
13) Call the REST service using node.js
14) Call TEIID from node.js
15) How to update tables from a VDB

This will be the "coming next" for this tutorial, if you are interested in some more integrations or further explanation please leave a comment.

TEIID tutorial 8.1 - Part 4

In the previous steps we created a relational model from a MySQL DB, we imported an XSD and created an XML model document based on that XSD and mapped the relational tables to the XML document.

Now we need to create the VDB so that all this work can be used in applications.

The Virtual DB will be treated as a normal DB and we'll connect to it with SquirrelSQL and issue a query to get back the XML document.

The process require to create in our project a new element that is the VDB database, give a name in the form and do not choose the models in the form, we'll to it later.

Now we find in the VDB view and we need to choose which are the models we want to publish: we need to remember that in a VDB we need to include all the models an which the model we'll query depend on.

If we choose the model we want to use TEIID Designer will automatically import all dependent models.

Also have to remember that if we make modifications in the models after we included them in the VDB we have to come back in this view and mark the checkbox that will be unmarked by default (this is a reason of many troubles).

After having included the models we have to save the VDB and we are ready to deploy it in a TEIID server.

Since TEIID is a module of JBoss AS weìll need to install TEIID into a clean JBoss installation.

We'll explain how to prepare TEIID to be able to deploy our VDB.

TEIID 8.1 tutorial - Part 3

In this part the target is to create the view model from the XSD to generate an XML document based on one of the XSD elements.

I choose bookListing.

You create a new Model and choose "XML" instead of "Relational"; obviously it asks you the XSD to use and you choose the correct one from the project.

Now you got a new "orange" model (a view model) with an object of type XML document.

The problem is how to fill the XML with relationa data.

The mapping tool comes to help, but you have to know well your DB and the XML you want to generate.

In this easy example you have to fill the book element and the author one: this means two mapping and the second depends on the first since the author depends on the book (in this XML).

So there goes the mapping, not too complicated (you could also write down the query by hand, but the tool is usefull).

I'm just not able to force it to choose "object" conversion instead of "string", but in any case is possible to fix it by hand.

Remember to save before invoking the tools otherwise an annoying popup will remind you that you have not saved.

In the end you have mapped a relational DB to an XML document with ease ... and now what?

Well you can query it with SquirrelSQL or publish it as a web service ... but before you have to do the most important thing create the Virtual DB.

But this is stuff for the next part of this TEIID tutorial.