RE: CLRLIB QRPLOBJ?? (was Re: IPL regains storage.......but)

Alexi:

The following is a quote from a normally totally reliable IBMer on technical 
issues from just over a year and a half ago:

"Because OS/400 may use QRPLOBJ to replace its own programs, the primary 
consideration is to avoid deleting programs which the system may be using. Most 
typically this would be a concern for *IMMED PTF applies."

The statement was made outside of any support channel because the subject was 
raised informally; therefore, there's no reason to mention who said it.I raised 
the issue then because I recalled warnings from IBM in earlier releases and 
couldn't recall any specifics. It is perfectly possible that the statement is 
obsolete for any release of OS/400 commonly in use today.

However, I have a couple specific concerns.

First, there is no doubt that system objects do get placed in QRPLOBJ under 
some circumstances. Second, a primary use of QRPLOBJ is to hold objects that 
may be in use and therefore cannot be directly deleted without 'unpredictable 
results'; by moving the objects to QRPLOBJ (even including renaming the 
objects), resolved pointers can continue to be used by active processes. This 
use can be seen by examining call stacks of jobs that are using such objects. 
An additional, secondary use is to hold objects that might need to be recovered 
quickly.

As long as system objects may end up in QRPLOBJ, I will always be wary of a 
CLRLIB. This is simply because they show up in there and I can't know why they 
weren't directly deleted.

Beyond that, even if there is no specific instance of potential trouble in 
current implementations of OS/400, I'm aware of no guarantee that a brand new 
PTF or OS/400 release that I apply tomorrow won't be a problem. If QRPLOBJ is 
used by Rochester developers, I'm going to respect standard usage. I have to 
assume that the object should not be deleted outside of the basic use of 
QRPLOBJ which allows for objects to be recovered up to the next IPL.

I easily agree that objects I can identify from my own development work are 
candidates for deletion whenever I choose because I chose to put them there. 
But I sure ain't going to advise anyone to do a CLRLIB without investigation. I 
have to recommend that the only objects to delete are ones you can identify.

If you want to make a flat recommendation to clear it at any time, be my guest. 
You probably have better sources than I have. But I hope you'll also supply a 
published reference in official documentation. I'll gladly take up the practice 
if IBM officially condones it. Given contradicting indications from IBMers I 
respect, I think it's worth resolving.

Tom Liotta


On Mon, 27 August 2001, "Alexei Pytel" wrote:

>
> I've had to do some investigation.
> I was wrong in my previous post.
> The bottom line is PTF process moves data to QRPLOBJ only if it is not
> supposed to be used anywhere on system.
> So as far as system itself is concerned it is safe to clear QRPLOBJ
> anytime.
>
>     Alexei Pytel
>
> "The better is the worst enemy of the good"
>
>
> ----- Forwarded by Alexei Pytel/Rochester/IBM on 08/27/2001 11:53 AM -----
>
>                     Alexei Pytel
>                                          To:     midrange-l@midrange.com
>                     08/25/2001           cc:
>                     06:26 PM             From:   Alexei 
>Pytel/Rochester/IBM@IBMUS
>                                          Subject:     RE: CLRLIB QRPLOBJ?? 
>(was Re: IPL
>                                          regains storage.......but)(Document 
>link: Alexei
>                                          Pytel)
>
>
>
>
>
>
>
>
>
> When PTF is applied - say a program object is replaced - the old object is
> moved to QRPLOBJ.
> In this way jobs which were using old program will continue to use it.
> Deleting a program when some job is executing instructions from this
> program, will cause the job to fail.
> This is why PTF process uses QRPLOBJ.
> It is safe to clear QRPLOBJ at IPL time, because there are no jobs using
> any object from QRPLOBJ.
>
> Earlier in this thread Tom Liotta said that you should be careful what you
> remove from QRPLOBJ - very true.
>
> Best regards
>     Alexei Pytel
>
> "The better is the worst enemy of the good"
>
>
>
> _______________________________________________
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@midrange.com
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/cgi-bin/listinfo/midrange-l
> or email: MIDRANGE-L-request@midrange.com

--
Tom Liotta
The PowerTech Group, Inc.
19426 68th Avenue South
Kent, WA 98032
Phone  253-872-7788
Fax  253-872-7904
http://www.400Security.com


___________________________________________________
The ALL NEW CS2000 from CompuServe
 Better!  Faster! More Powerful!
 250 FREE hours! Sign-on Now!
 http://www.compuserve.com/trycsrv/cs2000/webmail/