I don't know of any overt sign of flash copy invocation, but for me there is 
strong circumstantial evidence: lightning fast volume copy. We periodically 
migrate maintenance via full DSS volume copy. Done this for years. Then one 
time the copy job ran in only seconds instead of minutes. I thought for sure 
something had failed. Nope, all was A-OK. It happened that the source and 
target volumes were in the same DS8* DASD subsystem with the flash copy feature 
installed, so (I was told) flash copy got invoked automatically. When we copy 
between DASD subsystems, it still takes minutes. 

BTW I have a narrower view of 'online/offline' than Joel does. I take 'online' 
to be an MVS status involving OS control blocks. A volume is either online or 
offline on that basis. The trouble with bringing in 'reachable' is that it 
introduces a scale of 'accessibility' with many nuances. Is the device 
physically connected at all? If it is connected, is it genned to a chpid? If it 
is genned, is the chpid in the access list for that LPAR? If it is not in the 
list, can the IODF be updated to include it? These questions all cloud the 
issue. So, a volume is either online or offline to the OS at any given moment. 
Duplicate volsers are not allowed to be OS-online concurrently. No other 
restrictions. 

.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler 
SHARE MVS Program Co-Manager
323-715-0595 Mobile
626-302-7535 Office
[email protected]


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Joel C. Ewing
Sent: Saturday, December 03, 2016 8:18 AM
To: [email protected]
Subject: (External):Re: FLASHCOPY QUESTION

On 12/03/2016 07:58 AM, esmie moo wrote:
> Gentle Readers,
>
> I have a question about FLASHCOPY using  ADRDSSU and COPY.  In my first 
> example I was told that the job is using FLASHCOPY.  However I do not see any 
> evidence.
>
>  COPY FULL INDYNAM (SDB000) OUTDYNAM (MDB000) DUMPCONDITIONING           
> ADR101I (R/I)-RI01 (01), TASKID 001 HAS BEEN ASSIGNED TO COMMAND 'COPY ' 
>
> >From my understanding it is performing a full volume copy of a source 
> >volume.  If my understanding of the parm DUMPCONDITIONING is correct, it 
> >(DUMPCONDITIONING) allows to make a copy of the source volume and while 
> >keeping the target volume online.
>
> I have done a full volume copy of a source to target without the 
> DUMPCONDITIONING parm and it worked - the volume was fully copied (see 
> below).  Am I missing something? Aren't both examples doing the same thing?   
>                                  
>                                        
>   COPY  INDDNAME(DASD1) OUTDDNAME(DASD2) -                   
>   ALLDATA(*) ALLEXCP COPYVOLID PURGE                         
> /*                                                           
>
> However I do not see any evidence of FLASHCOPY being invoked because the FCNC 
> parm is not present.
>
> Here is a my second example:
>
>  COPY FULL IDY(PROM70,3390) ODY(SNAP63,3390) DUMPCOND FCNC- 
> ALLDATA(*) ALLEXCP CANCELERROR PURGE FCTOPPRCPRIMARY -     
> FCSETGTOK(FAILRELATION)       
>
> In the above example FLASCHOPY is being performed because of the FCNC.  
>
> Could you correct my understanding of all 3 examples?
>
> Thanks.
>
>
If you do a full volume copy without DUMPCONDITIONING and with COPYVOLID, that 
means when the copy is complete the target volume will now have the same VOLSER 
as the source volume and dss will force it offline, as you can't have two 
volumes with identical VOLSER online to
the same MVS system.   If your goal was just to make a point-in-time
copy on DASD, that may be sufficient.  But if your goal was to then back up 
that target volume to  removable tape media, you would not be able to do that 
with dss from the same system because the volume is offline because of the 
duplicate VOLSER.  It also means that if you for some reason needed to IPL at 
that point, the system will complain about finding two volumes with the same 
VOLSER and there could be some confusion about which of the two volumes should 
be placed online and the wrong one could be chosen. 

With DUMPCONDITIONING the original different target volume VOLSER is preserved 
so both volumes can be online to the same system and there is no later 
confusion about which is the original and which is the point-in-time copy.  DSS 
is smart enough when restoring from a tape dump of the volume with the 
DUMPCONDITIONING copy to also restore the original VOLSER (this used to require 
either an VTOCINDEX or VVDS dataset on the volume to supply the correct 
VOLSER).  

There are other vendor utilities that will allow one to dump an offline DASD 
volume with a duplicate volser to tape, but this seems to me a perversion of 
the meaning of "offline", which was intended to mean access by the system was 
impossible and that the device could safely be taken physically offline for 
hardware maintenance.

Assuming there is still a dss FASTREPLICATION parameter, it used to default to 
"PREFERRED", which meant if flashcopy wasn't available for some reason, dss 
would default to an ordinary copy but still do the COPY command successfully 
(just m-u-c-h slower).  If you wanted the copy to fail if flashcopy were not 
possible, you had to explicitly request FASTREPLICATION(REQUIRED).  It should 
be obvious from the time required for the COPY command to complete whether 
flashcopy has been used:  a fraction of a second for the dss COPY command to 
complete with fastcopy, versus minutes without fastcopy.
    Joel C. Ewing


-- 
Joel C. Ewing,    Bentonville, AR       [email protected]


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to