GraphDB Users Ask: How Do I Change the Configuration of an Existing Connector?

ONTOTEXT ANSWER:

Mistakes are a natural part of software development. We have the tools to assist you, but as changing configurations is a somewhat advanced procedure, we haven’t exposed it from the workbench. You can still change some configurations – you would just need the right SPARQL query.

As you probably already know, you can create connectors via a SPARQL query. In fact, this is what happens under the hood even when you use the workbench UI. That’s why it’s so easy to extract the query when you are making a connector.

Making a Lucene connector. The “View SPARQL Query” button is at the bottom left

You can also pick up the SPARQL query of any connector you have already created. Then, all you need to do is to change the command from createConnector to updateConnector.

Updating a connector

Now, this comes with a pretty big caveat: you can only update some of the connector settings! Why is that? Well, when a connector is created, GraphDB builds internal data structures and listeners. So, a change in a field or the type of the entities being indexed leads to a lot of internal changes. The changes are numerous and would necessitate a full index rebuild. At that stage, you are not updating a connector, you are rebuilding it – and there’s already capability for that.

So, what can you change? Many of the settings which are not critical to exactly what data is being indexed, but to how it’s being accessed. Namely:

• For Elasticsearch:
  • elasticsearchNode
  • elasticsearchClusterSniff
  • elasticsearchBasicAuthUser
  • elasticsearchBasicAuthPassword
  • bulkUpdateBatchSize
  • bulkUpdateRequestSize
• For SOLR:
  • solrUrl
  • bulkUpdateBatchSize
  • solrBasicAuthUser
  • solrBasicAuthPassword
• For Kafka:
  • kafkaNode
  • kafkaSSL
  • kafkaProducerConfig
  • kafkaCompressionType
  • kafkaPlainAuthUsername
  • kafkaPlainAuthPassword
  • bulkUpdateBatchSize
  • bulkUpdateRequestSize
  • kafkaPropagateConfig