'Give deployer complete control over Trino additionalCatalogs'

[mikej888]

In relation to monolithic_mk8s_example/1-vm-setup-and-deploy.yaml (originally from DARE-SeRP-Dev-Deployment Version: 43f688f (Thu Sep 14 17:15:42 2023 +0100) main branch).

I updated the appTrino additionalCatalogs minio catalog name in teleport-values.yaml to be edinburgh (a name that might be more usable for a researcher than minio. When I requested ArgoCD rereploy the Trino coordinator, the Trino pod pod/app-trino-coordinator-.... failed with:

2023-10-19T15:45:04.632Z	ERROR	main	io.trino.server.Server
No factory for connector 'hive_hadoop2'.  Available factories: [hive,
ignite, delta_lake, memory, kudu, blackhole, kinesis, redis, accumulo,
gsheets, elasticsearch, jmx, postgresql, redshift, sqlserver,
localfile, example_http, tpch, iceberg, mysql, bigquery, prometheus,
mongodb, pinot, tpcds, hudi, oracle, mariadb, druid, trino, cassandra,
clickhouse, kafka, singlestore, atop, raptor_legacy, trino_thrift,
phoenix5]

This seems to be because DARE-Trino's charts/trino/values.yaml defines:

additionalCatalogs:
    minio: |
      connector.name = hive-hadoop2
      ...

If there is no minio catalog in any overriding values, then this default catalog remains. So, in my case, I had edinburgh and minio catalogs.

Could values be updated to provide an empty additionalCatalogs block and/or could the charts/trino/templates/configmap-catalog.yaml be updated to replace this block with that from a deployer? I think any deployer should have complete control over the catalogs deployed and their names.

[mikej888]

Closing as fixed by fix(chart): set default extra catalogs to empty (#70).