Indexing is provided by a Spring boot application called the Elasticsearch connector. This application contains two main components that build and maintain the index in Elasticsearch.
Live Indexing: Metadata, and Content and Permissions from Alfresco Repository are consumed using ActiveMQ messages so they can be indexed in the Elasticsearch server. The information created and updated in the Alfresco Repository is not immediately available in Elasticsearch, because it takes time to process the messages coming from the Alfresco Repository. The previous Eventual consistency approach, based on transactions and used for Solr deployments, has been replaced by this new approach based on ActiveMQ messages.
Re-indexing: Indexing the information of a pre-populated Alfresco Repository or catching up with Alfresco Repositories that have missed some ActiveMQ messages is provided by the re-indexing component. Metadata and Permissions from the Alfresco Repository is retrieved using a direct JDBC connection to the Alfresco Database. Note: Only PostgreSQL is supported. The re-indexing application also generates content indexing messages in ActiveMQ in order to get the content indexed. It may take some time to process all these requests after the re-indexing application has finished.