TSG has been eagerly awaiting the release of Alfresco One Enterprise 5.1.1, which landed on July 20th. We’ve had a few weeks to play around with the latest release and this post will summarize our thoughts.
Why so much excitement about a patch release?
Longtime TSG blog readers may be wondering why we didn’t blog about the 5.1 release sooner, since it was released way back in March 2016. The reason is this bug: MNT-14543. Essentially, if your model utilizes d:content properties, any updates to documents of this type (ex: a property change) caused the document to drop out of the Solr index, making the document impossible to search for. Since TSG uses a d:content property in our models to handle document renditions, this issue made Alfresco 5.1 a non-starter for us. See this post on working with versions in Alfresco for more information on why we use a content property, and especially the “Technical Explanation” section for our technically inclined readers.
We’re happy to see that the issue was fixed in the 5.1.1 release. According to Alfresco, this fix was applied to the 5.0.4 release as well. If you’re an Alfresco customer on versions 5.0 or above, definitely look into upgrading, especially if any of your content models utilize a custom d:content property.
Notable Improvements in Alfresco 5.1
For full details, see Alfresco’s documentation for all of the new features in Alfresco 5.1. The items that TSG is most exited about are:
Smart Folders
Smart Folders are a new feature in Alfresco Share that allow you to display information in folders based on metadata, not physical location. See Alfresco’s release documentation on Smart Folders for additional details. TSG has been pushing our clients to break the “file system folders” habit in ECM for a long time. We feel that it is much faster for users to access content via search and/or “smart folders”. Smart folders are especially important in case management solutions (ex: insurance policy and claim). It’s good to see that Alfresco Share is now adding Smart Folder capabilities for these types of systems.
Model Manager
Before the 5.1 release, all alfresco models were defined in XML files. While functional, there are two main drawbacks to the XML approach:
- Updates Require Restart – being XML files in the bowels of the Alfresco file structure, any changes to these files require a restart of the entire Alfresco system.
- User Friendliness – beyond needing to write XML, which isn’t terribly difficult, administrators have to know all of the possible options from experience or documentation. With the XML, there’s no user interface guiding them through all the possible options when building a content model.
In the 5.1 release however, a new Model Manager is available in Alfresco share that lets administrators build content models in a user friendly GUI. This includes everything you would expect – types, aspects, properties, and constraints. Read more about the model manager here.
While this is a step in the right direction, one thing that was disappointing is that if you have existing models that are structured using the XML approach, they are not accessible in the model manager. TSG is hopeful that this capability is added in a future release.
Solr Sharding
In previous releases, Solr could only have one index. While this is fine for small and medium deployments, for very large deployments this is a problem since the Solr index would grow too large and performance is impacted. Solr sharding allows the index to be split up into multiple parts, and indexes even can be located on separate machines. This update is a very welcome addition, especially for customers that need to store a large amount of data in Alfresco. Read more about Solr sharding here.
Final Thoughts
Overall, TSG feels like the 5.1 release added a lot of new features, most on the back-end side. There were a few additional updates to Share not mentioned above as well which will certainly be welcome for clients that rely on it as their primary Alfresco user interface. With the 5.1.1 release, we’re happy to see that the Solr indexing issue was fixed and we can now start utilizing it on our client deployments.