Re: RCLSTG on IPL?

On 22-Nov-2013 12:47 -0800, David Gibbs wrote:

Does anyone know if there is a way to force the system to run a
RCLSTG while it's IPL'ing?

I've got a situation where an internal system has some object damage
... the guy who knows anything about the box is unavailable and I'm
nowhere near the system.

If the damaged objects are known, they can be deleted with the corresponding DLTxxx command; optionally preceded by data recovery.

Note: Reclaim Storage (RCLSTG) does *not* correct hard\physical damage, neither full nor partial. The request will merely effect what is called /damage notification/ whereby an object's already-detected [or in some cases whereby the /navigation/ of the object, but unlikely navigation of any data, may detect some] damage, then a message in the ranges of CPF8100 and CPF8200 will be issued to *SYSOPR. After the reclaim, DSPLOG MSGID(CPF8200 CPF8100) can be reviewed to find the objects that need to be deleted; optionally, after data recovery... noting, that the *only* recover for damage is to delete the damaged [portion, e.g. the sub-object as composite, of the] object. Any objects that are damaged, but for which that damage is not yet detected, the object-damage will not be /notified/ by the reclaim; excepting, as noted, the /navigation/ of the object detects the damage during the reclaim request, for which the MCH16xx precedes the CPF8xxx logged to the joblog and to QSYSOPR.

I don't even know if it's got a 5250 console or a HMC ... and nobody
in that office knows where the console is.

At one time there was a capability to define a request to be invoked [via a Data Area] within the SCPF job [after the system jobs start, but before coming out of restricted state], but AFaIK that nonsense was retracted; well, at least the *DBXREF invocation could not be invoked.

The [better IMO] option exists, to use the dedicated-batch [restricted-state batch] feature of End Subsystem; i.e. the Batch Time Limit (BCHTIMLMT) to have the request run for an allotted amount of time.