Fast, repeatable provisioning
Turn a multi-hour manual provisioning ritual into one parameterized, observable flow.
A platform engineer who stands up test instances
The trigger. “Stand up an ART instance with the optional products my solution needs, on the ESM it needs.”
The old way
Provisioning runs 32 minutes with no optional products and up to 3h39m with the full set. It is a long sequence of manual steps. You provision the base USI, IPL to RACF, run ansible prereqs, transfer product binaries and restore datasets, and configure RACF/ACF2/Top Secret per product. Db2 alone is ~17 minutes on RACF. Then there are several IPLs, you deploy QA binaries, and you IPL to the target ESM. The process is opaque, so it is hard to tell whether it is stuck or working. It is easy to get wrong, and it is redone from memory each time.
The Flow way
The same sequence as one parameterized, observable flow:
Set Variables (ESM, optional-product list, HLQ) →
Provision base USI (Action: ART CLI) →
Install optional products (Action/Zowe: transfer binaries + restore datasets) →
Configure ESM (Action) → Deploy QA binaries (Action) →
IPL to target ESM (Action) → Verify IVP (Zowe / Action) →
Log ready + Notify (Google Chat)
You pick the ESM and product list as inputs. Each long-running step logs its progress, so it is never a black box. The flow is identical across teams instead of living in any one person’s head. Pair it with the release-testing flow to tear the instance down automatically when testing finishes.
Payoff
| Old way | Flow way | |
|---|---|---|
| Effort | Hours of manual, sequential steps | One run, parameterized |
| Visibility | Opaque - “is it stuck?” | Per-step progress logged |
| Consistency | From memory, per engineer | Same flow every time |
| Choosing ESM / products | Re-plan each build | Inputs on the flow |