Hi all,
I am curious to know how you handled change management when rolling out changes to cluster access modes (single user, shared) applicable to UC, especially when there is a centralized platform team offering Infrastructure Deployment services via DevOps for around 20+ projects building their use cases in Databricks Lakehouse.
I am trying to decouple the changes to cluster access modes, and roll-out those even before the UC upgrade starts, so the projects can deploy their Databricks jobs/workflows in "Single User" access mode already in Non-UC enabled workspaces (like they do currently on high concurrency, table acl enabled clusters with external hive metastore). Even though I notice "Single User" access mode in Non-UC workspaces, it seems I cannot use it against the cluster policy of existing table acl enabled clusters with external hive metastore.
If this is not possible, then the option I think of is having a flag in the CI/CD code that signifies which environment (DEV/QAS/PRD) is UC enabled, and accordingly deploy jobs/workflows accordingly with appropriate cluster access modes.
