Managed elasticsearch node instance? about dropwizard-elasticsearch HOT 5 CLOSED

@msteinhoff Basically because it wasn't needed until now and because it's a "discouraged" way to connect to an Elasticsearch cluster (e. g. Elasticsearch 5.x will make that much harder and provide a "proper" HTTP client library for Java).

This being said, feel free to provide a pull request.

from dropwizard-elasticsearch.

Ah I see.

My use case for this is an embedded Elasticsearch node that also stores data and performs indexing in the dropwizard service, without any clustering whatsoever.

It works for my case, but I don't know if Elasticsearch recommends this. I try to create a PR, still have to figure out how to make my integration generic enough.

from dropwizard-elasticsearch.

@msteinhoff Since clients and full-fledged Elasticsearch nodes are a different thing (and should be handled differently), I suggest creating a class similar to ManagedEsClient (ManagedEsNode maybe?) which takes its own configuration (EsNodeConfiguration maybe?) and creates a data-storing Elasticsearch node.

from dropwizard-elasticsearch.

Yep, thats exactly what I have done already.

But I am not sure why this project does not follow the config factory approach recommended by dropwizard. I now have a EsNodeFactory (similar to the DataSourceFactory) which can build() a ManagedEsNode instance.

from dropwizard-elasticsearch.

FYI:

I started adding my changes in a local clone last week but noticed that I still use ES 1.x in my project (and the customer has no plans to change that). This addon currently depends on ES 2.x and there seem to be breaking API changes between 1.x and 2.x. So while I can add my changes, they won't do much for me.

Besides that, Elastic dropped support for embedded mode: https://www.elastic.co/blog/elasticsearch-the-server

With that in mind, I'll close this issue.

If someone wants to use my code, feel free to send me a mail.

from dropwizard-elasticsearch.