Re: 7.3 CHKPRDOPT fails after latest set of group ptf's

On 13-Sep-2016 14:45 -0600, Rob Berendt wrote:

Put on some group PTF's on numerous lpars on one of my Power 8's.
So far I'm batting 1,000 on them getting the following
CHKPRDOPT PRDID(*OPSYS)
...
*SQLPKG QSQLPKG2 in QSYS not found for product 5770SS1 option *BASE
release V7R3M0. CPD3DC8
Exit program processing found error in product.
Exit program processing failed.
Product 5770SS1 release V7R3M0 option *BASE load 5050 not correctly
installed.
...
Errors found by CHKPRDOPT. CPF0C20

The object DOES exist. It's the second reason:
-- When the Check Product Option (CHKPRDOPT) command is used, the
object is found, but the object description does not match the
product load.

DSPOBJD
Should not show:
Allow change by program . . . . . . : YES

The [most] relevant detail regarding the required "match" to the Product Load (PRDLOD) data would come from the Display Object Description (DSPOBJD) having requested to include DETAIL(*SERVICE), and for which the object description should show [on an IBM i 7.3 system]:

Licensed program . . . . . . . . . . : 5770SS1 V7R3M0

That "Allow change" setting of Yes vs No is merely a side-effect of the object being created on the system rather than having been restored from the install media as part of the OS install; the actual value of that attribute should be entirely unrelated to the error msg CPD3DC8 issued for that SQL Package object identified during the Check Product Option (CHKPRDOPT) request.

I restored from another lpar. Save file / ftp thing...
Reran CHKPRDOPT. Pass with flying colors.

That is a functional recovery for the *SQLPKG object [program] type, but not an option for programs generally; the system entry point table would need to be updated also, for most *PGM objects in QSYS to be able to be /recovered/ with that method.

PMR: 91188 082 000

Note: Create date of 'bad' object is on a Friday about noon.
2016-08-26.
Odd. Wondering if I ran some API to recreate it because I was having
some other issue?

The generic object auditing I expect would reveal what program and job created the object in a T-CO entry, possibly other(s) such as a T-OM entry and\or T-OW would appear just after that. With the qualified job name, possibly there is a joblog; otherwise what other audit [and history log] entries were logged for that job might be somewhat informative of what the job was doing.

Then again, maybe it is the PTFs. We'll see if PTF's corrupt it on
other lpars with a pre-ptf create date much earlier.

The following topic thread "Subject: Three-part names and QSQCLIPKGx SQL packages", for which below is given a link to the most recent message], may be of interest to review; that message similarly mis-identifies the "Created By User" as the required "match" to the Product Load (PRDLOD):
[https://groups.google.com/d/msg/comp.sys.ibm.as400.misc/56NmV6J1-qc/PDSRU5XDBwAJ]