Tuning and Configuring Server
Combine is designed to handle sets of metadata small to large, 400 to 4,000,000 Records. Some of the major associated server components include:
store Records and their associated, full XML documents
store Transformations, Validations, and most other enduring, user defined data
store transactions from Validations, OAI requests, etc.
used for indexing mapped fields from Records
main engine of field-level analysis
the workhorse for running Jobs, including Harvests, Transformations, Validations, etc.
used to send and queue Jobs to Spark
the GUI
for long running tasks that may that would otherwise prevent the GUI from being responsive
includes deleting, re-indexing, exporting Jobs, etc.
Given the relative complexity of this stack, and the innerconnected nature of the components, Combine is designed to be deployed via an Ansible playbook, which you can read more about here. The default build requires 8g of RAM, with the more CPU cores the better.
This part of the documentation aims to explain, and indicate how to modify of configure, some of the these critical components.