customer Security

ACF2 security scan, without the mainframe tax

Turn an hour-long, four-system relay into one click - and make 'your toolkit was out of date' impossible to hit by accident.

Michal, systems programmer at a Tier-1 global bank

The trigger. “Run the ACF2 security configuration scan (MRISEASC) on this LPAR and get a clean result file ready to upload for our Security Essentials assessment.”

The old way

Michal runs the full four-system relay. He downloads the toolkit in the web UI, FTP-uploads the two ACF2 transmit files with hand-set allocations, and RECEIVEs them in TSO. Then he rewrites the MRISEASC job card and the HLQ/SYSPLEX variables in the JCL editor, submits the job, and digs through the return codes to work out what happened.

There is also a trap underneath all of it. He reuses the toolkit already on the LPAR without realizing it is back-level. So even a “successful” run can produce a file the web UI later rejects, which sends him into a support ticket and a redo from scratch.

The Flow way

He opens the MRI Toolkit: ACF2 Security Scan template, or he types his “I need to…” into the generate screen. The graph runs these steps:

Prepare MRI Toolkit (Action) → Upload SEA.CNTL.XMIUpload SEA.LOAD.XMI (Zowe) → Render install JCL (Action) → Submit install JCLPoll install status (Zowe) → Submit MRISEASC scanPoll scan status (Zowe) → Download scan spool. It also has a built-in failure branch to Download spool (on failure) for when the scan return code is not clean.

He sets his connection and HLQ once and hits Play. The toolkit is fetched fresh and pinned across every step, so there is no stale copy or half-upgraded mismatch. The result file lands locally and is ready to upload. If anything fails, the diagnostic spool is already downloaded.

Payoff

Old wayFlow way
Tools touched41
Manual steps~121 (Play) + one-time config
JCL / allocation editingBy hand, every runTemplated, filled once
On failureManual log huntSpool auto-downloaded
Version driftSilent fail, ticket, full redo (days)Fetched + pinned per run; fix pushed via hub update