2

Is there any feature provided by "ml-gradle" wherein I can deploy replica forests topology with Availability zones in mind?

Or any API by MarkLogic to know which node is in which Availability zone so that I can define my own logic for the forest replicas deployment.

Dixit Singla
  • 2,540
  • 3
  • 24
  • 39
  • What are your actual requirements? Like how does the availability zone that a host resides in matter to your policy for creating forests? https://github.com/marklogic-community/ml-gradle/wiki/Creating-forests describes the approaches available for creating forests. You always have the "payload-driven" approach as a fallback in case the set of properties don't meet your use case. – rjrudin May 13 '20 at 12:26
  • 2
    I'm looking at http://docs.marklogic.com/REST/GET/manage/v2/hosts/[id-or-name]/properties and there is a "zone" property. So ml-gradle could definitely be enhanced so that it supports a property for creating forests and replicas based on zones. But that gets back to my first question - what are your requirements? Specifically, when you use the properties that ml-gradle supports today to create forests and replicas, what problems do you run into with how they're created and distributed across hosts? – rjrudin May 13 '20 at 13:32
  • 1
    I am looking for this as well. I can see on the Hosts page, that MarkLogic is aware of which AZ the hosts are in, but when I do a deployment including replicas, I end up with replicas in the same AZ as the host, which is not what I would like. My cluster is a managed cluster with 2 hosts per AZ across 3 AZs. – H. Garrow Nov 28 '22 at 18:30
  • https://github.com/marklogic-community/ml-gradle/issues/635 – Mads Hansen Dec 01 '22 at 18:53

0 Answers0