Shutting down and checking an IPL
A good IPL starts with a clean shutdown: subsystems stopped in order, then JES2, then the system. After start-up, proof matters more than impressions: D IPLINFO shows which members and volume were used, D A,L shows what is running, and the run book's checks confirm the change took effect.
Shutting down in the right order
Stopping a system is the reverse of starting it. Applications and subsystems are stopped first so that they close their data cleanly, then JES2, then the system itself. Automation usually drives this from a single shutdown command, but operators watch it and deal with anything that does not stop.
| Order | What stops | Why |
|---|---|---|
| 1 | Online work: CICS regions, then Db2 and MQ | Transactions finish and data is committed and closed cleanly |
| 2 | Batch: let jobs finish, or stop them at a safe point | No job is left half-way through updating data |
| 3 | JES2 ($PJES2, once no work remains) | Spool and checkpoint are left consistent for a warm start |
| 4 | In a sysplex: remove the system from the sysplex | The other systems know it has gone and take over its resources |
After start-up: prove it
D IPLINFO
IEE254I 06.09.40 IPLINFO DISPLAY
SYSTEM IPLED AT 06.00.07 ON 10/18/2026
RELEASE z/OS 03.01.00 LICENSE = z/OS
USED LOAD01 IN SYS0.IPLPARM ON 0A82
IEASYS LIST = (00,01) (OP)
IODF DEVICE 0A82
IPL DEVICE 0A80 VOLUME MPRS01D IPLINFO answers the questions that matter after a planned IPL: which release is running, which LOADxx and IEASYSxx members were used, and which volume the system came from. (OP) shows that the operator supplied system parameters at the prompt. D A,L then shows the address spaces that are running, and D SYMBOLS shows the system symbols in use.
When start-up goes wrong
- A parmlib error: a member with a typing mistake can stop start-up or bring up the system without something it needs. The fix is usually to IPL again with corrected or previous members.
- The wrong volume or load parameter: the system may come up at the wrong level and look healthy. Only the checks catch this.
- A disabled wait state: start-up stops and the HMC shows a wait state code for the systems programmer.
- A subsystem does not start: the system is up, but part of the service is not. Automation reports it, and the owning team is called.
Which console command shows the LOADxx member and IPL volume used by the current IPL?
Show a hint
D followed by one word.
Show the solution
D IPLINFO.
Examples are for learning. Run commands and jobs only on a system you are authorised to use, such as a training or test system, and never on production without approval.
Common mistakes
JES2 should be stopped only when no work remains. Stop the subsystems and let batch finish first.
A system IPLed from the wrong volume looks fine. D IPLINFO is the evidence.
A cold start loses every job and all output on the spool. It is never a way of tidying up.
What you will see at work
- Run books for planned IPLs end with a list of checks and who to tell, so the change record can be closed with evidence.
- If a planned IPL goes badly, the backout is usually an IPL from the previous system residence volume, also in the run book.
- Practise the full sequence in the Command Center IPL drill.
Key terms
Check your understanding.
Take this lesson's quiz and save your progress. Free.