-
Notifications
You must be signed in to change notification settings - Fork 0
For Developers: Kubernetes
The rails application is deployed as a Deployment
in kubernetes (https://github.com/psu-libraries/researcher-metadata-config/blob/main/charts/researcher-metadata/templates/deployment.yaml)
Delayed Job is a kubernetes Deployment
( https://github.com/psu-libraries/researcher-metadata-config/blob/main/charts/researcher-metadata/templates/delayed_job.yaml)
Rake tasks are configured in helm in the cronWorkflows
property, and driven by Argo Workflows
- Non-Prod https://argo.dev.k8s.libraries.psu.edu/cron-workflows/researcher-metadata-qa
- Prod https://argo.k8s.libraries.psu.edu/cron-workflows/researcher-metadata
Postgres is hosted on VMs (postgres1qa, postgres1prod). Databases and roles are created via helm/crossplane (https://github.com/psu-libraries/researcher-metadata-config/tree/main/charts/researcher-metadata/templates/externalPostgres)
Kubernetes manifests are generated via helm. The helm chart is stored in the researcher-metadata-config repo (https://github.com/psu-libraries/researcher-metadata-config/tree/main/charts/researcher-metadata)
Cronjob docs will go here
ArgoCD/Rancher docs will go here
We use Argo Workflows to Restore a production database into a preview, or QA environment.
The Restore workflow will grab a list of database backups from s3, scale down the rmd deployment, restore the file into postgres, and then scale back up the deployment.
- https://argo.dev.k8s.libraries.psu.edu/workflow-templates
- fill in your namespace (e.g researcher-metadata-qa for qa)
- Choose the
restore-database
task - Click "Submit"
- On the pop up dialog, choose "Submit"
- Once the "Choose File" task shows up, click on it and choose "Resume"
- Select your file from the Dropdown and click "OK"