Jonathan Ball wrote:
> Is there an easy way to locate all objects that were saved specifying
> STG(*FREE)? We had a job that ran out of control when it tried to do
> DSPFD on such a file. The job wouldn't end, couldn't be killed, and
> was consuming 20% and more of CPU. IBM eventually got in and killed
> it ("Piranha" connection through our HMC - some kind of black magic),
> but I'd like to identify all such files and get rid of them pre-
> emptively.

One idea is DSPOBJD to an OUTFILE and query it. The field ODOBFR has a
value of '1' if storage is freed.
dspffd QADSPOBJ

The Storage attribute returned by the List Objects (QUSLOBJ) API gives
the same info.
http://publib.boulder.ibm.com/infoce...is/quslobj.htm

--
Karl Hanson

Karl Hanson wrote:
> Jonathan Ball wrote:
>> Is there an easy way to locate all objects that were saved specifying
>> STG(*FREE)? We had a job that ran out of control when it tried to do
>> DSPFD on such a file. The job wouldn't end, couldn't be killed, and
>> was consuming 20% and more of CPU. IBM eventually got in and killed
>> it ("Piranha" connection through our HMC - some kind of black magic),
>> but I'd like to identify all such files and get rid of them pre-
>> emptively.
>
> One idea is DSPOBJD to an OUTFILE and query it. The field ODOBFR has a
> value of '1' if storage is freed.
> dspffd QADSPOBJ

I don't mean to appear greedy, but what I was hoping for was to discover
some storage-related command that would just bring up a display of them
based on a parameter in the command. Doesn't seem there is one. I'm
familiar with the object information APIs, so I suppose I could write my
own.


>
> The Storage attribute returned by the List Objects (QUSLOBJ) API gives
> the same info.
> http://publib.boulder.ibm.com/infoce...is/quslobj.htm
>
>
> --
> Karl Hanson

Jonathan Ball wrote:
>
> I'm familiar with the object information APIs, so I suppose
> I could write my own.

The /storage freed/ object attribute for a database *FILE object has
no meaning. The suspended\freed attribute is tracked to each member,
because any one member may have had its storage freed while any number
of other members have not.

Regards, Chuck

Jonathan Ball wrote:
> Is there an easy way to locate all objects that were saved
> specifying STG(*FREE)? We had a job that ran out of control
> when it tried to do DSPFD on such a file. The job wouldn't end,
> couldn't be killed, and was consuming 20% and more of CPU. IBM
> eventually got in and killed it ("Piranha" connection through
> our HMC - some kind of black magic), but I'd like to identify
> all such files and get rid of them preemptively.

It would seem appropriate to ask for the PTF that is preventive to
the loop; or that the PTF be created if one does not exist.

Regards, Chuck

CRPence wrote:
> Jonathan Ball wrote:
>> Is there an easy way to locate all objects that were saved
>> specifying STG(*FREE)? We had a job that ran out of control when it
>> tried to do DSPFD on such a file. The job wouldn't end,
>> couldn't be killed, and was consuming 20% and more of CPU. IBM
>> eventually got in and killed it ("Piranha" connection through our HMC
>> - some kind of black magic), but I'd like to identify all such files
>> and get rid of them preemptively.
>
> It would seem appropriate to ask for the PTF that is preventive to the
> loop; or that the PTF be created if one does not exist.

There is one, and I've obtained it. But obtaining the downtime
necessary to apply it in an IPL is harder in our organization than
carrying the Hope diamond out of the Smithsonian, so I'd like to do
something to try to eliminate the risk, until we can put the PTF on
probably some time early in the year.

Jonathan Ball wrote:
>
> There is one (a PTF), and I've obtained it. But obtaining the
> downtime necessary to apply it in an IPL is harder in our
> organization than carrying the Hope diamond out of the Smithsonian,
> so I'd like to do something to try to eliminate the risk, until we
> can put the PTF on probably some time early in the year.

FWiW, No PTF for that code should require an IPL. And there are
unlikely to be any requisites that would require an IPL to apply.

Regards, Chuck