Select Page

FAQs - Upgrading Alfresco content services

Zia has compiled a list of the most frequently asked questions about upgrading Alfresco Content Services. From deployment to databases, your questions can be answered below. Don’t see a solution that works for you? Scroll down to contact Zia today.
 
  • image/svg+xmlimage/svg+xml
    How do I get started with upgrading Alfresco Content Services?
    Get familiar with the documentation. Upgrade information is available on the Alfresco docs site. Depending on the version you are currently using, you might have to perform a full re-index of your SOLR indexes. If you are upgrading, you might also plan on migrating data from on-prem to cloud. The migration plan as recommended by Alfresco will need to be tailored for each cloud provider.
  • image/svg+xmlimage/svg+xml
    How do I choose the right Alfresco deployment method?
    Check out our blog article Alfresco Content Services Production Deployment | Part 1: Review of Deployment Methods. We published another article in that series recently: Part 2: Deployment with Docker Compose. We’ve also done a series about Getting Started with Alfresco and Kubernetes.
  • image/svg+xmlimage/svg+xml
    How do I test for Alfresco upgrades in lower environments?
    This may seem obvious to some. It is surprising how many customers who attempt their own upgrades are not aware that their production license grants them licenses for some lower environments. It is relatively trivial to spin up a blank instance with your existing version, maybe add your customizations and upgrade in order to get a sense of things. Ideally you would actually copy your production data to a lower environment and perform all of this testing with all of that data in place.
  • image/svg+xmlimage/svg+xml
    Do I need to upgrade my Alfresco customizations?
    In some cases it is not necessary to recompile or adjust customizations when upgrading. Especially when those upgrades are between minor versions. Having said that, it is always a good idea to validate that existing customizations still work with the ACS version you are upgrading to. In other cases there are changes that will require some rework of code; e.g., for folks upgrading from older versions of ACS to recent versions. Jobs will likely need a little work and any custom transformation will likely need quite a bit of work. The Alfresco Extension Inspector can help when you have complex customizations.
  • image/svg+xmlimage/svg+xml
    What should I consider when upgrading from On-Prem to a self-hosted Cloud?
    Consider performing the upgrade and the move to cloud as separate initiatives. While it is possible to upgrade and move to cloud at the same time, this can add significant complexity and risk to each project. If you plan to use the AWS S3 Connector or the Azure Content Connector, don’t forget that you’ll need to update the prefix on the content_url column from “store://…” to “s3v2://…” or “azb://…” in the database.
  • image/svg+xmlimage/svg+xml
    Can I switch databases while upgrading Alfresco?
    Customers frequently ask if they can switch databases as part of an upgrade. Switching database technologies is minimally supported at this time. In the past this was not supported at all. There is no tooling provided for such a migration, so you will have to find a solution outside the Alfresco product stack. Our general advice is to strongly consider not switching database stacks. If it’s absolutely necessary to perform such a switch, we recommend not attempting to do this at the same time as an upgrade.
  • image/svg+xmlimage/svg+xml
    Why is Alfresco Search Services not starting after changing JDK version?
    Use the latest version of Alfresco Search Services 2.0.3. This fixes an interesting bug where SOLR cannot start when the JDK version has 4 parts to it. https://alfresco.atlassian.net/browse/MNT-22911. This version of search services is certified with Alfresco 6.2.1 and above. If you are forced to run an older search services version, consider installing a JDK that does not have 4 parts. For example, 11.0.14.1+1 will break Alfresco Search services < 2.0.3 In this case, go with 11.0.15+10 or 11.0.14+9 where the JDK version has only 3 parts. Note that Alfresco Search Services 2.0.3 will only work with Alfresco Content Services versions less than 7.2 with SSL communication. The none/http method was removed. The new secret method does not exist in the earlier versions of ACS.
  • image/svg+xmlimage/svg+xml
    How do I disable auto-suggest?
    The Solr auto-suggest feature is enabled in one place in Share. It provides auto-completion based on the search term the user is typing. This has been found to block indexing of new content when you start SOLR for minutes to hours. It leaves files on disk that you have to clean up manually. This takes extra CPU and RAM every hour and provides a feature that most customers won’t notice. Setting the following property to false in alfresco-global.properties and the solrcore.properties files in both alfresco and archive cores turns this feature off:
    solr.suggester.enabled=false
    More details on this can be found in the article written by Angel Borroy.
  • image/svg+xmlimage/svg+xml
    Why are my SOLR results intermittently empty?
    For Alfresco 7.1 & 7.2, we have noticed SOLR queries that disappear without giving results or a clear error message. It turns out that, if you get a timeout when communicating with SOLR, it raises an IOException that is just rethrown as a RuntimeException with no message: https://github.com/Alfresco/alfresco-repository/blob/787d860c7b7c29e09e40ebb2f4c5ce0a2e70142e/src/main/java/org/alfresco/repo/search/impl/solr/SolrQueryHTTPClient.java#L597 So if something similar happens to you, we've been increasing the following settings in order to fix it:
    • solr.http.socket.timeout=10000
    • solr.http.connection.timeout=1000
    The default values were:
    •  solr.http.socket.timeout=2000
    •  solr.http.connection.timeout=500
    You might want to tweak the values based on your own environments.
  • image/svg+xmlimage/svg+xml
    How Much Content Not Yet Indexed; Search Services 2.x
    Because metadata and content are indexed in parallel (meta first then content), it can appear that indexing is finished when all the metadata has been indexed. If you need to know how much content is left to be indexed, Angel Borroy came up with a nice Solr query.

Do you have a technical question for our team?

Didn’t find the answer you were looking for? Contact our team today and we’ll get back to you within the next couple business days about your question.

Pin It on Pinterest