The latest release (7.14) of the elasticsearch Ruby client adds some code that prevents the gem from working with “unsupported” versions of Elasticsearch. “Unsupported” in this context means not only “OpenSearch,” but also the Apache2-licensed versions of Elasticsearch for some reason.
As a hosted search provider that supports both OpenSearch and Elasticsearch, we’ve already had customers writing in after getting exceptions in test and PR apps, thanks to the latest elasticsearch-ruby release. Our advice has been for users to pin their apps to the 7.13 release of the gem for now.
We are in need of a longer-term solution, namely an Apache 2.0-licensed Ruby client. I noted in the announcement that there are plans for doing exactly this.
How can we best help in that effort? Or at the very least, speed up the timeline for availability?
Welcome to the forum. Glad to hear that you’re willing to help. Right now the team is in the process of making these repos free from potentially problematic trademarks and other similar activities, so they are marked as private. However, you can still contribute but there is a few extra steps to get you access. I’ll PM about early access (or you can wait until they are public/public - I don’t think it will be too terribly long).
We are also trying to switch to OpenSearch and appreciate the commitment for gem opensearch-ruby. However our code also uses gem elasticsearch-model, gem elasticsearch-rails and I’m finding nothing on the internet about ruby applications that utilize those how to make them work with OpenSearch. My guess it they would also have to be forked, and code re-worked to use OpenSearch instead of ElasticSearch language throughout our app codebase? It seems elasticsearch-model has a dependency to elasticsearch which leads to that unsupported message. Sticking with the <7.14 versions by elastic is not a very long term solution. So re-opening this since I can’t find info on the internet about anyone else running into the same issue, in case someone else search the same. Would really like to jump on the OpenSearch bandwagon if anyone has a solution. cc: @searchymcsearchface @robsears