How to properly scale up Search nodes?

Issue

• How to properly scale up Search nodes instead of deleting the Search service and deploying it again?

Environment

• Liferay PaaS
• Liferay DXP 7.4+

Resolution

• Following the Liferay Cloud Upgrading to a High Availability Subscription documentation, which also describes the procedures to scale up search services, what you need to do is:

  (… to) update the search service to increase its number of instances it uses by increasing the scale property in your project repository’s search/LCP.json file. Increment the value to the next odd number of instances to prevent issues with the service starting (generally from 1 to 3 when upgrading to the High Availability subscription):

  { "kind": "StatefulSet", "id": "search", "image": "liferaycloud/elasticsearch:6.8.13-4.1.1", "memory": 8192, "cpu": 8, "scale": 3, "podManagementPolicy": "Parallel", ... }

  WARNING

  In order to ensure your search service’s instances can connect to each other properly, increase the scale property to an odd number of nodes, and ensure that the podManagementPolicy property is set to Parallel.

  So the advisory in this WARNING must be followed exactly to prevent connection issues.