Preparing for Deployment

There are a few things that must be done before the deployment happens and while the builds of deployable artifacts are ongoing.

  1. Create a Jira ticket for the cycle build configuration work. You will need a new ticket for each deployment site.
  2. Prepare the configuration for the cycle build in the git repositories listed in Deployment Configuration Repositories using the Jira ticket above as the branch.
    1. Use the site specific file/directory: Summit, NTS.
    2. Update the cycle build tag.
    3. Update any changes to CSC configurations including launch command-line.
    4. Add new CSC/applications as necessary.
  3. Work with build team during the build process to ensure schedule and resolve any encountered problems: Summit.
  4. Create cycle build Confluence page (Software Upgrades) with versions of core packages and any operational changes.
    1. Announce page once all base artifacts are built.
  5. Announce the deployment schedule on the slack channel: Summit, NTS.
    1. Use the release_announcement.py script from vanward to craft the announcement.
    2. If you want to work with the System Principles for DM and Camera machines, make sure to inform them you will require their help standing down services.
  6. Coordinate with SQuaRE to make sure that a new nublado with the current XML/SAL will be available for the deployment day.
    1. Make a PR for the site specific science-platform configuration here: https://github.com/lsst-sqre/phalanx.
    2. Edit the appropriate configuration file: Summit, NTS.
    3. Notify SQuaRE when the PR is ready to merge.
    4. Syncing the cachemachine app will take place during the deployment.

Deployment Configuration Repositories

As noted above, following repositories contain the configuration as code.

Deployment Helper Repositories

The following repositories contain helper scripts that aid in the deployment process. Specific uses for each repository are handled within the deployment documentation.

Site Specific Variations