Tag Archives: .net

Untangling the Sitecore Search LINQ to SolR queries

2 Replies

Problem

It can be very difficult to identify why you do not get the search results you expected from Sitecore Search, but there is a simple way to help untangle what is going on.

Solution

It is possible to see the query that Sitecore generates and sends to SolR and then use the query on the SolR instance to see what data is returned to Sitecore.

This is such a huge help when trying to understand why your queries do not work!

Step 1 – Find the Query that was sent to SolR from Sitecore

Sitecore logs all the queries it sends to SolR in the standard sitecore log folder, look for files named Search.log.xxx.yyy.txt .

Step 2 – Execute the query in your SolR instance

Go to your Solr instance, and use the core selector drop down to select the index your Sitecore Search query is being executed against.

Select Query, from the menu

Then paste the query from the sitecore log, and you can see the result that is returned to Sitecore.

This has helped me a lot, so I hope this helps others untangling their search results using Sitecore Search 🙂

 

 

 

Sitecore Helix – Supporting Integrations

1 Reply

This blog will outline how it is possible to identify, separate and isolate dependencies on external systems, by introducing an Integrations module group. See here for more information about module groups in Helix.

But why do we need a Integrations module group?

Any feature that requires integration to achieve its purpose, will introduce additional challenges relating to stability and additional system knowledge, than a standard feature layer module.

Stability

Helix is built on a number of principles that help deal with stability, the Stable-dependencies principle and Stable-abstractions principle for more details see my blog post.

Features that rely on 3rd party systems are by nature more unstable than any other feature/foundation modules. As it is not usually possible to control when external systems change, upgrade or fail. Therefore, quality assurance, test, and automated deployment for the website cannot protect against this type of change or failure.

System knowledge

Working with a 3rd party system, such as CRM, ERP, Marketing Automation, DAM, SolR, Custom API, etc. requires additionally system specific knowledge. So it is a good idea to use abstractions to hide the system specific complexities for several reasons:

  1. The web team might not have the specific system knowledge.
  2. The web team should not be distracted by the additional complexities of all the integrations.
  3. The team responsible for implementing the integration may have limited Sitecore knowledge and should not be distracted by the rest of the web solution.
  4. It is good practice to separate modules by their responsibilities by splitting the presentation and the retrieval of data from the external system.

Integrations Module Group

The intention/purpose of the Integrations module group is to clearly define which modules have a dependency on an external system and ensure they are only responsible for that integration to the external system.

Example

This solution is responsible for selling a wide variety of courses. The customer has its own custom course catalog API and a complex legacy enrollment system.

The following diagram shows the module architecture for the solution.

In the foundation layer the following modules were introduced to provide abstractions. If you are not familiar with IoC and Abstractions see my earlier post.

  • Course Catalog
    • Defines the abstraction/system agnostic logical data module for the course catalog.
    • Acts as an abstraction between the website and Custom Catalog API.
    • It helps to focus on the ideal model that supports the business objectives.
  • Enrollment
    • Defines the abstractions to support the process of initiating attendance on a course at a specific school and shopping cart.
    • Acts as an abstraction between the website and the enrollment legacy system.
    • It helps to focus on the ideal model that supports the business objectives.

In the Integrations Group in Features layer the following modules where introduced

  • EXT Course API
    • Responsible for getting data provided by the EXT Course API.
    • Provide the implementation of the Course Catalog (foundation layer) abstractions.
    • Responsible for caching the course catalog, as the API only supports periodical batch retrieval.
  • EPOC Enrollment Management
    • Responsible for integration of the functionality provided by the EPOC Enrollment Management SDK.
    • Provide the implementation of the Enrollment (foundation layer) abstractions.

In the Features layer the following modules where introduced

  • Course
    • Responsible for the presentation of the course catalog, retrieved via the course catalog abstractions defined in the foundation layer.
  • Enrollment
    • Responsible for presentation and controlling the process of initiating attendance on a course and displaying the shopping cart, using the abstraction  defined in the foundation layer.

Additional Bonus

Once the integration code is isolated in a single module and only responsible for the integration, it is easier using dependency injection to achieve the following:

  1. Update external system, as the code to change is clearly defined and separated from the presentation and website logic.
  2. Provide the ability to support more than one version of an integration (i.e. different sites use different version)
  3. Move the integrations modules to an integrations platform, if it is the domain model for the customer business.

I hope this blog post gives you some ideas on how to isolate and remove the complexities introduced by integrations from your Sitecore solution, Alan

In my next post, I will explain why and in more detail how to isolate the SolR integration.

 

 

The RED dot is not Enough

Leave a reply

The classic check to ensure a test is running is that the red dot (see image above) on the optimization tab is shown, if so all was good. Unfortunately, that is not true, as it is possible that the red dot is shown, but in fact the tests for the page are not running. Sitecore are currently fixing this issue.

So how do we know if the test is running or not? You have to check that status window, for example below the test is NOT running, as the status window says “No Tests”. Thanks to Alec Orlov  from sitecore for this tip.
If the test is running, the status window will contain the estimated number of days for the test to complete, see below.

So what can cause this issue? Well in this case, the customer had started (Deployed) the test from the “Analytics Testing Workflow” see the image below.

It worked for some Sitecore versions and or with a solution specific patch, but it does not work in general. The workflow is in an internal Sitecore workflow, which should not be used. Please follow the official Sitecore documentation, to start your tests.

Solution

If you have a test that is not running, a common issue is related to the fact that the test item which is stored under /sitecore/system/Marketing Control Panel/Test Lab is not in the correct workflow state (must be in deployed state) and or is not published.

I hope this helps, Alan

Sitecore Commerce 8 powered by Microsoft Dynamics (SCpbMD) – Find the channel ID

1 Reply

Sitecore expects an integer value to identify the channel id, which is used to identify the online store to retrieve from the channel database.

The problem I faced was that the neither the Microsoft Ax Partner or the customer could give me the id, all they could find was the name which was a string value!

But fear not the following SQL statement which you can run against the channel database will show you all the channels and their assoicated id 🙂

SELECT TOP 1000 [NAME]
      ,[CHANNELID]
      ,[OPERATINGUNITNUMBER]
      ,[SERVER]
      ,[DATABASE]
      ,[ISPUBLISHED]
      ,[ISLOCAL]
      ,[CONNECTIONSTRING]
  FROM [AsyncClientDB].[crt].[STORAGELOOKUPVIEW]

Also ensure that channel is in fact published otherwise Sitecore can not use it, hope this helps Alan

channel id

How to filter AD roles or users using Sitecore’s LDAP module

2 Replies

Problem
Every time I have used AD for providing access to Sitecore, the active directory (AD) structure is crazy and recently I had a customer that had over 18000 roles, which made it difficult to assign roles and it killed the performance of the Sitecore client, as each user had at least 500 roles. Therefore Sitecore to evaluate the combination of a lot of roles to determine if they had read access or not.
18000I talked to the department responsible for the AD setup about changing and or creating a folder that only contained the Sitecore related roles, but this was not possible.

Initially I thought I would have to make own LDAP provider which derives from the standard provider, but I discovered this was not necessary as the LDAP module provides the functionality as standard.

Custom Filters
Custom Filter provide the ability to filter the roles and or users returned from the AD (see section 4.1 for full documentation).The custom filter uses the standard LDAP query syntax (see MSDN) to specify how the user or roles are filtered.

The following example ensures only roles, which contain Sitecore and or the special operations role; are imported into Sitecore. The Customcustom Filter

According to Sitecore documentation, both the User and Role provider must have the same CustomFilter, and that is why the (objectCategory=person) is added so all users are also imported regardless of their name.

I hope this blog post will help others using LDAP to control what roles or users are shown within sitecore.

C# Regions

1 Reply

Whilst reviewing a solution I came across the class in the image below, and felt the need for a rant about using regions.

Regions C#

If you need regions, you should consider if the class has more than one RESPONSIBILITY and or is too big! In this case it is responsible for everything and over 1800 lines of code:-(

So a tip, if you need to use the #REGION’s in your code. I would advise you take a few moments and ensure that the class does not have mixed responsibilities. If it does have more than one responsibility you should split it up into a number of smaller classes with a clearly defined responsibility.

I re-factored the class into the following classes:

  • 4 Service classes
  • 13 Factories
  • 13 Repositories
  • 13 Model classes.

In fact by the time I had removed all the code from the class I found that it was not required at all 🙂