On-demand jobs

This project also provides two sets of on-demand deployments of the OSO tools that are useful for executing more realistic observations against a real TMC deployment.

Note unlike the persistent deployments, these jobs deploy all the OSO applications to single namespace rather than simulating the three sites in three namespaces.

On-demand with real TMC in simulation mode

In the on-demand pipeline stage, the deploy-on-demand-with-real-tmc manual job will deploy the ska-oso-integration-umbrella chart to a temporary environment in the integration-ska-oso-integration-real-tmc namespace. This environment will be destroyed after 6 hours. The URLs for the applications deployed to that environment are those listed in SKA OSO Integration, but with the namespace changed to integration-ska-oso-integration-real-tmc.

This deployment includes the real TMC devices with simulations of the subsystems. This is closer to a real deployment and is useful for testing the OSO-TMC interface. The same real-TMC-with-simulated-subsystems pattern underpins the k8s-test-mid and k8s-test-low jobs used by the integration-ska-oso-integration project CI/CD pipeline.


../_images/real-tmc.png

On-demand with an external TMC

This environment can be used to control a real telescope deployment (for example from one of other the integration repositories) using OSO. It requires a telescope deployment with a Tango host available.

In the on-demand pipeline stage, the deploy-on-demand-with-external-tmc manual job will deploy the ska-oso-integration-umbrella chart to a temporary environment in the integration-ska-oso-integration-external-tmc namespace. This environment will be destroyed after 6 hours. Again the URLs for the applications deployed to that environment are those listed in SKA OSO Integration, but with the namespace changed to integration-ska-oso-integration-external-tmc.


../_images/external-tmc.png

This job requires variables to be set when running. In the GitLab UI, set

  • TANGO_HOST to the name of the Tango databaseds Kubernetes Service. This will be something like

    tango-databaseds.<TANGO_NAMESPACE>.svc.<CLUSTER_DOMAIN>:10000 where TANGO_NAMESPACE is the Kubernetes namespace that the Tango databaseds (and likely the rest of the telescope) is deployed to.

  • SKA_TELESCOPE to either ska_mid or ska_low.


../_images/pipeline-variables.png

ODA deployment and persistent data

These environments connect with the same persistent ODA instance that the main integration deployments described in the previous page use. This means entities from previous on-demand deployments will be available, as well as any created from the main deployment.

Dev branch deployments

Dev branches all have a job in the deploy stage that deploys an on-demand environment with the OSO-TMC Simulator. The namespace is ci-dev-ska-oso-integration. Data created in these environments will be lost on each redeployment.