Pipeline Deployments

This project provides a permanent deployment of the OSO tools for stakeholders and developers. There are two environments:

  • integration deploys from the main branch of ska-oso-integration and provides the latest versions of OSO software. Applications in this environment should be stable and compatible with each other, though there could be unforeseen issues as this is the first place the whole OSO suite is used together. Please report any erorrs to #help-oso.

  • staging deploys from a tagged released version of ska-oso-integration. This environment should always be stable and corresponds to a set of versions that are confirmed to be compatible, and can be picked up by the AIV teams.

Alongside the OSO applications, a simple TMC simulator (that is maintained by OSO) is also deployed so that the OET can execute simulated observations.

For deploying an environment to use OSO to execute real observations against a real TMC deployment, see the next page.

Namespaces

In production, some OSO tools will be deployed globally to the cloud, and some will be deployed at each of the telescope sites to run against either a Mid or a Low deployment. To simulate this, the persistent environments described above are split across three separate namespaces - integration|staging-oso-cloud, integration|staging-oso-mid and integration|staging-oso-low.

../_images/persistent-deployment.png

For example to define an SBDefinition for either telescope, access the cloud ODT deployment, then to execute that SBDefinition access the OET instance in the correct telescope namespace.

ODA deployment and persistent data

To ensure any data created in the integration environments is persisted, there is a long-lived deployment of the ODA in the <integration|staging>-oda namespace, controlled by a separate pipeline job. This means data should be available from previous deployment versions.