AWS Step Functions to Start/Stop the Staging Environment

Description

This repository includes code and Terraform manifests for an AWS Lambda and AWS Step Functions to reliably start and stop resources within the NUL RDC staging environment as follows:

Spin Down

• Scale all ECS tasks in meadow, avr, and arch clusters down to 0
• Perform incremental backup of avr and arch collections in Solr
• Scale all ECS tasks in fcrepo and solrcloud clusters down to 0
• Stop meadow-db and stack-s-db RDS instances

Spin Up

• Start meadow-db and stack-s-db RDS instances if necessary
  • Wait for both instances to reach available state
• Scale fcrepo ECS service up to 1
• Scale up solrcloud ECS services
  • 3 Zookeeper nodes
  • 4 Solr nodes
• Wait for the Solr cluster to report 4 stable, live nodes
• Restore avr and arch Solr collections
• Start meadow, avr, and arch ECS services

The Terraform manifests also create AWS EventBridge Rules to spin the stack up and down every weekday at 6am and 8pm Central, respectively.

Execution

To spin the staging environment up or down, execute the stack-s-spin-up-environment or stack-s-spin-down-environment Step Function, respectively, with the following payload:

{
  "solr": {
    "baseUrl": "http://solr.internal.rdc-staging.library.northwestern.edu:8983/solr/",
    "collections": ["arch", "avr"]
  },
  "rds": {
    "meadow": "meadow-db",
    "stack": "stack-s-db"
  }
}

There is also a staging-environment utility script on the staging bastion host that can be used to spin the staging environment up or down on demand.