Rolling version update deployment utilizing dynamic node allocation

Abstract

Methods are disclosed for dynamic node allocation for a server system that can automatically heal on failure—minimizing the need for static configuration—dynamically adjusting server resources to match load, and minimize end user wait times. The disclosed methods dynamically allocate nodes to increase capacity for a platform that accepts data queries. Additionally disclosed is a system for rolling version update deployment: workers maintain org lists of org-task-queues that they service; org-affinities between the workers and the org-task-queues require the workers to have access to local copies of immutable data sets to service org-tasks from the org-task-queues of the orgs that they service. A leader running on a worker implements a healing and balancing service that maintains worker redundancy, manages the workers' org-affinities to accumulate orgs on their respective org-lists. The leader implements messaging to the workers to update to a new software version and monitors completion of updates.