Miss Integration

3 Simple steps to Reindex Social Communities

Shelly Mao — Tue, 24 Jan 2012 04:39:40 +0000

Hello everyone!

This is my first post in 2012 but it will sadly.. to be my last post of the OpenText Web Solutions Product. I will soon become the newbie in the Content Server space, yes still in the OpenText space so watch out I am coming!

Here I want to quickly document how to configure Search Engine Settings using Vignette Collaboration , especially for those ones like me came from the old RedDot and LiveServer world, the admin console seems pretty …. complex!

Just a few steps, I am sure you will find them easy to follow.

Step 1

Log into Social Communities / Vignette Collaboration and browse to the Main Page. The URL is by default http://:8007

Click on the “System Administration” link

Click on the “Administration Console” link

Click on any of the “Services” links

Her you see a list of Services and one of them is the “Search Service”. Click on that and you will see the screen below

Step2

Now we are in the right place to kick off some actions. Before you go ahead of kick of the Reindexing action, let’s click on the “Search Engine Information” link down the bottom of this page. You will see something like this

Step 3

Once we are all happy with the config and ready to reindex, you have two options here. If you know the Remote Object ID of the content you’d like to reindex, you can just do the “Manual Object Indexing”; if you are unsure and would like to reindex all the objects, there is also an option for that which is “Data Based Indexing”.

Social Communities uses Common Search so you can start monitoring the log files and hopefully that just worked for you!

Best Approach for Managing CSS and JS

Shelly Mao — Mon, 14 Nov 2011 02:09:53 +0000

What is the best approach while implementing the content structure?

Should CSS and JS be removed from a project and kept separate from content?

Have developers and designers ever fought over how the files should be accessed?

Should files be stored and managed by IIS or CMS? Have you ever wondered how to make everyone happy?

Thanks to Graeme Heath, the Web Designer/Info Architect with 5+ years experience working with the content structure in CMS, here is the summary of our “best practice approach” in delivering site structure which we think is the best for now.

Over the last 2 years a lot of talk has focused on getting CSS and JS files out of a Management Server project and putting them into a separate publication target or website. The two main approaches were either linking the CSS and JS resources into to hard linked locations created by clever project variants/publication packages. The other method was to keep all the CSS/JS files in an external site and just link to them. (The first method was demonstrated in the Best Practice project that came with MS v10).

Both of these approaches did not suit us – the Best Practice approach was cumbersome for designers wanting to create or update CSS and JS – see above. (Pretty well every project we’ve done has required the designers to update and expand the CSS and JS files frequently). The external site approach loses the benefits of working with assets in Management Server – that of being able to selectively publish to different targets and move between environments.

For the last 6 months we’ve been using a compromise solution that has given us more flexibility but still kept all the assets under management, and allowed us to publish to the publication server/s via publication packages.

Here’s the approach in a nutshell:

Put the CSS and JS into asset manager folders.
Create a virtual directory that points to a parent of the CSS/JS asset managers.
For preview/SmartEdit mode, link directly to the main CSS/JS files via the virtual directory.
For published pages, use standard media items for the main CSS/JS files, and link the child assets (background images, fonts) from a dedicated text field.

The Advantages:

Designers will love you. They have their independence and power back. If you deal with highly strung designers, this is a bonus*. Designers will do extra changes for you without complaining. The project will be a success.
Simplicity – the CSS and JS files are left intact as created by the designer. Adding new CSS or JS files is straightforward. No need to add hundreds of image/media placeholders. Changes to the CSS and JS can be done externally via a third party tool.
Speed – compared to the traditional approach of using templates for CSS, this approach means a template does not have to be assembled and rendered for each stylesheet.
Ease of moving content to new servers – compared to using a dedicated external CSS/JS server location you have control over where the CSS/JS files are published to. When you copy a project you can change the targets and all the CSS/JS routing is configured via the publication packages.

Note that we have not implemented this for a multilingual site. If CSS or background images are language dependent then a dedicated CSS folder may be required for each language variant, and hence multiply the effort.

Some details:

You’ll need to put all the font files and background images into the CSS directory and its sub-directories. It’s good practice anyway and makes CSS much easier to package and move around between projects.
The designer can build and test the CSS in a Dreamweaver project (yeah, designers still love DW) and can set up remote server connections and push new files and versions to the virtual directory location. If you _have_ to use Managment Server for versioning then use check-in/out from the DW project location. We find that using file versioning on DW projects is easier to manage (using SSafe or SubVersion).
Use only one level of sub-folders as Asset Manager will ignore anything deeper.
Linking the CSS and JS was accomplished using the following render tags


                                                      <%stf_LocalServerAssetLocation%>/css/css-main.css


                 <%med_CSSMain%>

   ” />

Create a dedicated text field to hold the links to CSS background images and font files. After doing one sub-directory of images and looking at the source code you’ll realise there’s a simple pattern to the links and you use can a copy/paste sequence to create lots of links in minutes.
Source Code View:

CSS Files

css-data.css

css-location.css

…

Using Excel to generate the list of links will speed up the time taken to create them. If you’ve added new sub-folders then run Update Asset Manager from SmartTree before yiu try to connect to the new items.
If the publishing server uses Delivery Server, publish the CSS/JS files to Tomcat/Apache folders. These can handle sub-directories and the relative internal file links will stay intact.

(* Confession – I’m a designer too. On _very_ rare occasions I’m highly strung)

Graeme Heath

One of the many ways to achieve LDAP SSO using Delivery Server 10.1

Shelly Mao — Mon, 23 May 2011 05:47:01 +0000

Have you ever tried to Single Sign-On LDAP users into Delivery Server projects? Last week I was looking for a quick and easy method, with special assistance from Tim Davis (and thank you!), here I am documenting the steps to achieve this.

My solution can be explained in the simple diagram below:

As you can read from the diagram, I am using .NET to read the user name, format my own digest string, and then include hash value as one of the request parameters in the URL before getting redirected to Delivery Server by ISAPI. Once the request has traveled to Delivery Server, I can then use the username provided in the URL, compute and compare the hash values; if the hash values match then I can use dynament code to login the user automatically using the trusted mode.

Step 1 – writing .NET code
This one is easy, very straight forward; a few variables, one function, one redirect.

At the end of the Page_Load, I redirect the user request to a virtual directory I made up with the same name as the Delivery Server project name (just to make the URL user friendly).

Step 2 – Configure the lovely ISAPI

As many of you know, ISAPI is used to redirect traffic from IIS to Apache. Depending on the complexity of your project, redirect rules need to be defined in the ISAPI config file. I am not going into too much details about setting this up because there are many materials out there you can ask Mr. GOOGLE for.

Windows authentication is definitely required here.

A Quick Overview at where we are at so far

We have two parameters in the request variable, one is the hash value produced by my function in .NET, and the other one is the user id in clear text. We now need to confirm that the hash value is correct; since MD5 is a hash value and can’t be reverted back to it’s original string, we need to produce our own hash value using dynament code and compare the values that way. In order to get the hash value using MD5, a validation rule needs to be set up in Delivery Server, and this is what we will do in the step 3.

Step 3 – Set up Validation Rule on Delivery Server

Delivery Server allows customized rules to be defined via Admin Console.

In my dummy project, I created a rule using GetDigest with MD5 (SHA is also available)

Step 4 – Dynaments

Now we have the validation rule ready, we can start developing our dynament code.

1. Format the digest string

2. Read Token Value

I had some problems with the value being cached during testing but not appearing in the object cache listing; this is just a workaround for the value to be read on every page refresh.

3. Compare hash values

Here is where we compare the hash values and decide whether we can automatically log-in the user. This algorithm can be extended to check whether the user is already logged on, and included in your normal login XML.

That is it! I believe there are other ways to achieve the same result, so please feel free to provide some feedback and your experience

Troubleshooting Common Search with Delivery Server [Best Practice Project]

Shelly Mao — Wed, 20 Apr 2011 08:39:57 +0000

I went through the post blogged by Danny Baggs over and over again for about more than 10 times while setting up my Common Search server but it was just not working! There is nothing more frustrating than having the search working in Livelink but not in Delivery Server. I am new to the whole LiveLink interface thing and that didn’t help me either. If you are on the same boat and are having similar issues, I hope this post will assist a little.

From reading Danny’s post, we all know that we need to set up a shared folder and a Delivery Server connector to communicate with Common Search server. Here is the summary of what I’ve concluded after my 3 day’s tough time in regarding to how the integration works.

Delivery Server creates xml files which contain all the information about pages and project content, those files are stored under the shared folder for the Common Search Server to process; therefore, both servers need permission to write and read contents in this shared folder. The Common Search Server then indexes the files created by Delivery Server, creates a collection and sets up a search scope for each project . When a search query is performed, Delivery Server sends a request to the Common Search Server, the search query gets executed, and returned to the Delivery Server Server in the XML file format.

Easy, yeah?

So, here we set up the Delivery Server Connector, and I want to use my own Common Search user but not the Admin Account.

First tip, make sure the user has the “System Administration Rights“. This can be configured in the Common Search Interface.

The next thing is to kick off the indexing job. This is not hard, everyone with some Delivery Server Experience would know where to find the button.

Once the Search Engine Indexing job is started for the first time, you will find a few directories created automatically for us in both the Shared Folder and the Working Folder.

There is also a new Data Source created automatically in Common Search; this can be found by logging into the Common Search Administration Interface.

Once you’ve clicked on the Data Source with your project prefix, here the prefix is wsm and the project name is wsgbpp (Best Practice Project), you will find a few objects created in this Data Source Folder including the Data Flow Manager object.

The Data Flow Manager object contains a very important member called the XMLActivator object. This object is the key for this integration, it manages all the communication between Delivery Server and Common Search Server.

With this XMLActivator object. there are two things we need to do here.

1. Give it a schedule to run periodically so the new index files created by Delivery Server can be processed; and

2. Start the XMLActivator object.

That is pretty much it. You should now be able to test the Search Connector and get the following result.

If you are still having issues getting the results back to Delivery Server, it is always recommended to test the search functionality using the Common Search interface first. To do so, log in as the user into the Common Search, go into “Advanced Search”, choose the scope (you should be able to see the scopes available here), and perform testing.

This is what it should look like in the Best Practice Project -

That is it from me. Happy troubleshooting!

Installing Web Compliance Manager with Management Server (RedDot)

Shelly Mao — Wed, 20 Apr 2011 05:39:27 +0000

Web Compliance Manager is a separate module can be purchased to ensure that your web site meets corporate compliance without spelling errors. More marketing information can be found here .

This is a really simple install, and I am sure there are many people out there have implemented this without issues. I am documenting this more for my own benefits, so I don’t have to scratch my head again when doing the installs.

This time I will be installing a Pre Version, as you are probably aware, there are two versions “PRE” and “POST” for WCM. The “PRE” version checks the pages before they are published, while the “POST” version checks against the already published pages/sites. In this example, I will be using the PRE version, and will be installing on the Management (RedDot CMS) Server 10.1 using the Tomcat Server as the server engine.

Firstly, you need to get a license; (as always…) and secondly, you need to download the required software from the Knowledge Center.

Step 3

Decide the server engine you are going to use, and get the prerequisites installed. For example, here I need Java and Tomcat. The Installation files are structured very neatly which also comes with software required for prerequisites.

Step 4

Deploy the wcm.war file. If you are using Tomcat like I am, it is very easy. Simply copy the wcm.war file into the [Tomcat Install Dir]\webapps folder, and restart the Tomcat Service. Once the service is restarted, you will see the WCM folder created under the webapps folder.

Once the web service is deployed successfully, you should be able to browse to it by visiting the address

http://:/wcm/services

Step 5

Install the Web Compliance Manager Interface. During this installation, you will need to specify the URL where the Web Service is installed in the previous step.

Step 6

We are pretty well done here, now we just need to configure the Management Server. Note that the configuration here needs to be done for every projects require WCM.

If you are just using the Validation component, you will only need to configure the Project Variants; but if you are also configuring the SpellCheck component, you will need to configure the Language Variants as well.

WCM comes with a few pre-defined rules; however, customized rules can also be developed. ( For more information please check the install manual. )

Once these are configured, you should be good to go. Simply select a page and click on the options in the Action Menu. You should see the following results.

That’s all from me, good luck!

SeeUnity’s DM Integration for SharePoint – Overview

Shelly Mao — Wed, 30 Mar 2011 07:32:08 +0000

Recently I got my hands on some of the SeeUnity’s powerful Integration Components. SeeUnity has two components designed for SharePoint, the Enterprise Integration and the Archiving and Distribution Services. The first one allows users to access or push eDocs documents via the common SharePoint Interface, and the later one allows documents to be synchronized or linked (manually or automatically) between SharePoint and external repositories based on customized rules and field mappings.

There are many functionalities provided by these two components, but doesn’t matter how many you are configuring, there are two things need to be understood before getting your hands dirty.

Understand the number of applications required to be installed and
Decide where they should be installed.

It sounds easy, but for me this was the most difficult part. In most environment, there are at least 2 servers involved (given that the chance of having both the SharePoint Server and DM Server on the same box is extremely low); the DM server and the SharePoint Server. To make it even more complicated, some of the components can even be installed on a dedicated SeeUnity Server for performance optimization.

The SeeUnity’s documentations are fairly well documented, but not all the scenarios can be covered; and the fact that it is so flexible, it does add some complexity to the installation process.

The core applications for the integration are:

Core Integration Services

Administration Console

Web Parts

WebDAV Server

Protocol Handler or BCS Indexing (Security Trimmer optional)

Archiving and Distribution Services

Of course, there have been tough times in getting all these two communicate, but the components seem to be working stably after it is configured correctly. Phew~ At least

OpenText Social Communities Installation + Integration with MS SSO

Shelly Mao — Wed, 09 Feb 2011 08:35:43 +0000

I have to say, I struggled with this for a good few weeks. Since the Delivery Server 10.1 is out with the Social Communities included, we have been asked many times to demonstrate it to our clients. Potentially, huge money there. However, the support to get this up and working is very limited.

I will try and make this really short, so that I don’t bore you to death.

Step 1

As always, make sure you have the required licenses, assuming you already have a working instance of Delivery Server 10.1. There are three modules that you will need, which are

HTRE – HTTP REST Connector
PCON – Portlet Connector
PRDW – Interface to JSR 168 and IBM WebSphere portlets

The OpenText licensing department seems pretty efficient in issuing/renewing licenses these days, which is good to see.

If you can’t get those required licenses, you will not be able to continue.

Step 2

You will need to install Common Search which now comes free with the Delivery Server. (Make sure you have the license key too) You will need to do this before installing the Collaboration Server.

Step 3

We now need a Collaboration Server. Collaboration Server can be downloaded from the OpenText Knowledge Center along with the installation manual. You will need to follow the instructions carefully, as well as setting up the correct permissions for the indexing folder/s.

A successful install should get you the screens like below.

: Collaboration Login Screen

: Collaboration Interface Screen

Step 4

After getting those pre-requisites ready, we can now import the transport package which comes with the DS installation file, named “social-community-integration-tp-xx.x.x.xxx.zip”. Make sure there is no error messages during the import, the process should be painless. You should now have 2 demo projects imported, portlets set up, as well as HTTP REST connectors.

Step 5

Phew~ All we have to do is just configuration on the Delivery Server now. You will need to set up the Collaboration Server Details, verify Portlets Details, as well as the user details.

The working copy should look something similar to these images below.

Blog Listing – here you can add and manage blog posts

: Blog Listing

Blog Management – here you can manage settings and permissions through the interface

: Blog Management

Forum and Wikis are similar.

Forum Listing – here you can see all the content

: Forum Listing

Forum Create New Post – here it allows users to create new posts

: Forum New Topic

If you want to go even further….

Step 6

You know what? We can bring those blogs into the Management Server using the Single-Sign On Functionality! Look at the screen shot below, pretty neat hey!

: CMS Integration

Bear in mind that those sample projects are just to demonstrate us what can be achieved with Delivery Server, but not a final product. I am sure going to hassle our Web Designer to beautify its look

Who said RedDot CMS is going away? OpenText MS 10.1 will convince you!

Shelly Mao — Mon, 11 Oct 2010 02:59:21 +0000

A lot of people are scared that RedDot OpenText Management Server might be going away since being acquired by OpenText a few years ago; it doesn’t seem to be the case by looking at the number of versions released after that! I recently installed the most recent version of RedDot OpenText Management Server 10.1 and it definitely has impressed me in many ways.

Best Practice

First of all, the new Best Practice sample project is heaps better than the up-and-away project. It does take a while to set it up but once it is set up everything you need is all documented there!

For those people who are experiencing difficulties getting it up and running, here is the tip – make sure that the default CSS files are published into the default website folder under the inetpub dir in your C drive.

Supported Operating Systems

10.1 is the first version of CMS which supports Windows 2008 and R2; however, it requires a bit more hardware requirements than 2003 if you want it to perform fast.

Integration Matrix

It was interesting to see that the eDocs integration version specified in the readme file is 4.5.1.12; this is the version didn’t work with version 9. (See the 10.1 Readme file for more integration version details.)

Documentation

Documentation wise, I have to say that OpenText has put in a lot of effort in getting it up to the standard. The online help has totally changed and it is pretty helpful.

New Asset Panel and the extended Drag & Drop Feature

Oh the new asset panel totally amazed our technical CMS graphic designer; it has made our lives so much easier to enter and update website content. However, I have been getting issues when trying to upload images L

Many more

There are heaps more new features in 10.1 that you can find in the OT knowledge base.

Who said RedDot is going away? Just show them the Best Practice if you hear it again!

eDocs/HummingBird and CMS Integration

Shelly Mao — Wed, 06 Oct 2010 07:45:21 +0000

Many people asked me about the integration between OpenText eDocs (former Hummingbird) and OpenText MS (former RedDot). This is going to be an ongoing post, since the area is too large to include every detail in a post. I will just try my best here, but please bear in mind that every environment is different and do not expect the integration to work the first time after your install!

First of all, you have to ensure that you obtain the correct version of the integration piece for your environment. Here is the compatibility matrix, the version I have been working with is eDocs v6 and CMS 7.5, CMS 9, CMS 10.1.

Got the right version? Now we can start.

On the eDocs Document Management Server, make sure that the followings are installed and required components are created with sufficient permissions.

IIS (eDocs WebTop)
SOAP Toolkit
.NET Framework 2
COM+ components are created (detailed in the eDocs DM installation manual)
ensure that RedDotUser is created, this is used for connecting to CMS server (detailed in the CMS Integration installation manual)
ensure that RedDotRole is created, and whoever is used for anonymous access is added to this group
configure DCOM (detailed configuration is specified in your CMS Integration installation notes)
Double check that DCOM security settings are set correctly (this is often the cause of miserable issues)

On your CMS Server, do the following:

Install SOAP
Install XML Parser
Install DM Integration piece (follow the integration installation manual)
Here is an important part – Not every CMS install is always successful. I have come across a couple of times that the web services required for integration were not generated during the initial install. So, make sure that services exist under the directory [%Root%]\RedDot\CMS\ASP\WebService. If not, there are files under the support directory which allows you to regenerate the web services.
Confirm that the web services are up and running by checking the WSDL file in IE.

By now, if you are not feeling frustrated and dizzy already, you should be able to fire up the webtop on the CMS box. Make sure that you test the webtop access “ON” the CMS box. It is sometimes necessary to test the access under the CMSUSER account.

How do I know whether it works?

It is simple. Just look for these icons after logging onto your webtop via CMS.

Now, you can start editing your CMS templates. Oh before you do that, a HummingBird Folder needs to be created first in your Asset manager. (The steps are well documented in the Integration User Manual.)

Troubleshooting

You can turn on the Logging functionality on the eDocs server by modifying the flag on your DM server.

Once the integration is configured correctly, everything should just work hunky-dory. One more thing – if your system is going to be used in a large organization, you might need to work out how many concurrent users is required and change your connection pool settings. or you will encounter random error messages and connection issues.

Hope you find the tips above useful