Re: V7R1 SQL Testing Issue - RCLSTG without console access

[CRPence <CRPbottle@xxxxxxxxx>]

> "Richard Schoen" wrote:
>
> <<SNIP>>
>
> It shuts down all subsystems, runs RCLSTG and restarts subsystems
> QCTL and QINTER.
>
> Probably best to do a PWRDWNSYS restart after running RCLSTG if
> you were  to use this on a production box.
>
> It's not pretty but it works.
>
> <<SNIP BCHTIMLMT() CLP to effect reclaim *dbxref>>

  There would be little value to perform a PWRDWNSYS after 
reclaiming the DB cross-reference.  In fact, just the opposite.  The 
system jobs QDBSRVXR & QDBSRVXR2 [or iASP equivalents] that perform 
the work of the reclaim run asynchronous to the requester of RCLSTG, 
and so powering down stops the processing, thus making those jobs 
continue where they left-off when they were interrupted by the PwrDwn.

  FWiW: For a RCLSTG SELECT(*ALL), there was long ago a functional 
benefit to powering down just before that request, with a side 
effect of some impact to its performance [similar to saves after IPL 
versus delaying saves until after normal system utilization over 
several hours].  A design change eliminated the pwrdwn as a 
pre-requisite to obtain the alluded benefit, but having added an 
improbable but potentially highly-negative side effect, which AFaIK 
has never been corrected.

Regards, Chuck