In over my head with gremlin / Java

[unimaginativeusername]

I am trying to connect to a remote database:

Docker

services:
  arcadedb:
    ports:
      - 2480:2480
      - 8182:8182
      - 5432:5432
    environment:
      - JAVA_OPTS= -Darcadedb.server.rootPassword=playwithdata -Darcadedb.server.plugins=GremlinServer:com.arcadedb.server.gremlin.GremlinServerPlugin,Postgres:com.arcadedb.postgres.PostgresProtocolPlugin
    image: arcadedata/arcadedb:24.6.1
    container_name: arcadedb-public
    networks:
      - reverse-proxy
      - arcadedb

networks:
  reverse-proxy:
    external: true
  arcadedb:

Setting up the connection:

 ArcadeGraphFactory pool = ArcadeGraphFactory.withRemote(graphDbConfig.getHost(),
                    graphDbConfig.getHttpPort(),
                    graphDbConfig.getGraphName(),
                    graphDbConfig.getUser(),
                    graphDbConfig.getPassword());

Getting the graph:
return pool.get();

Using it:

 ArcadeGraph g = ConnectionPool.getGraph();
        List<? extends Vertex> accounts = g.traversal().V().hasLabel("Account").has("email", email).toList();
        g.close();

This works, as long as I create a DB called "graph". I have need now for using multiple schemas at the same time and would like to access them as above.

Calling the traversal results in

 Servlet.service() for servlet [dispatcherServlet] in context with path [] threw exception [Request processing failed: java.util.concurrent.CompletionException: org.apache.tinkerpop.gremlin.driver.exception.ResponseException: The traversal source [auth] for alias [g] is not configured on the server.] with root cause
org.apache.tinkerpop.gremlin.driver.exception.ResponseException: The traversal source [auth] for alias [g] is not configured on the server.

I suspect I need to change a server config? Is it at all possible to use multiple databases in parallel?

(Everything works in Studio; I can access both databases via Postgres driver, too, using the same config. Both databases - graph and auth - were created through studio on top of the Beer example.)

My setup is a bit Frankenstein'ed, but I think I have provided everything that matters for this issue; I use postgres solely to get Flyway to cooperate during boot.

Edit:
In my installation, I copied gremlin-server.properties into gremlin-server-new.properties and included the path to the new database; I then extended the graphs-entry in in gremlin-server-yaml to include graph: ./config/gremlin-server-new.properties

It boots with:

2024-12-15 16:55:11.435 INFO [ServerGremlinExecutor] A GraphTraversalSource is now bound to [g] with graphtraversalsource[arcadegraph[new], standard]

2024-12-15 16:55:11.438 INFO [ServerGremlinExecutor] A GraphTraversalSource is now bound to [graph] with graphtraversalsource[arcadegraph[new], standard]

Access fails with The traversal source [new] for alias [g] is not configured on the server.

Edit2.
From the documentation:

By default, the database "graph" will be available through the Gremlin Server. You can edit the database name or add more databases under the Gremlin Server by editing the file config/gremlin-server.groovy

I guess I'm in the wrong place, but I can't see what to change here?

[unimaginativeusername]

I found a solution after some more trial and error:

in gremlin-server-new.properties , rename the variable from graph to new.

// define the default TraversalSource to bind queries to - this one will be named "graph".
globals << [graph : traversal().withEmbedded(graph), new : traversal().withEmbedded (new)]
// define the default TraversalSource to bind queries to - this one will be named "g".
globals << [g : traversal().withEmbedded(graph), n : traversal().withEmbedded (new) ]

gremlin-server.groovie
include the new variable as above, i suspect the redundancy will not help.

[robfrank]

Do you think we should update the docs and/or the code somewhere?
Thanks for your contribution!

[unimaginativeusername]

I do, it boils down to "everything, everywhere"
I've been using OrientDB before, am trying to get Arcade up now (and see that the community edition of orient is being resurrected, too?) - and it is outright frustrating to try and get what should be simple things to run, and I have pretty much given up on understanding them.

I love the idea behind Arcade, and loved it when Orient came out, too. Janus and NEo4J seem to be too big and cumberosme to get into, really. Arcade pretty much just runs, but I get stuck as soon as I want to implement anything production worthy.

Case in point: Multiple databases under a server,

I have no idea how I should access the database (do I use a Pool? A Traversal Provider? What even are these things?) And there are hardly any sources anywhere else, because nobody seems to be using Graphs.

It seems like I have to jump between technologies to find working examples for different things. I would love to see complete code examples that demonstrate and explain all of the options, in all of the supported languages.

(I am using Gremlin, and that I think is the one exception - because it is standardized, once I have my Traversal I can work with it and find relevant documentation.)

Ideally, there would be a handful of demo projects that allow users to set up a working environment so that they can build and explore, rather than struggle with the basic setup. High availability? Different users, with different rights? Can I force the DB to require a schema? etc. etc.

Sorry if this is out of scope for this issue. Yes, I think there should be some instruction on how to get a single, named DB to work, too, But there is a bigger problem with the documentation....

[lvca]

Gremlin Server was originally intended as one graph per server. If you want to have multiple graphs you have to play with gremlin configuration and create multiple instances. The best would be exposing ArcadeDB databases at runtime under the Gremlin Server, but I couldn't find a way to do so.

The easiest way is to just use ArcadeDB HTTP Server, not Gremlin server. If you still want to use Gremlin as query language, you can execute Gremlin statements via ArcadeDB server.