GRS star is vastly superior in performance to ring mode. We saw a huge
improvement in converting even what amounted to a one-member sysplex. Some
points to consider.

-- While GRS ring can be configured to cover almost any (sub)set of systems
in a shared complex, the scope of GRS star is *exactly* one parallel
sysplex; no more, no less.

-- The star structure ISGLOCK is absolutely vital to the entire sysplex; if
it breaks or is unreachable for any reason, every member in the sysplex
will wait state and be unIPLable in star mode until ISGLOCK is once again
viable.

I've never run a parallel sysplex with only one CF. With only one CEC on
the floor, I would try mightily to run two ICF LPARs. If necessary, the
second (backup) LPAR can run in DYNDISP mode so that it uses very few
cycles normally but is available for failover or occasional off-shift
structure rebuild in order to do maintenance on the primary LPAR.

.
.
JO.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
626-302-7535 Office
323-715-0595 Mobile
[EMAIL PROTECTED]


                                                                           
             Alvaro Quintupray                                             
             <[EMAIL PROTECTED]                                             
             SSA.CL>                                                    To 
             Sent by: IBM              IBM-MAIN@BAMA.UA.EDU                
             Mainframe                                                  cc 
             Discussion List                                               
             <[EMAIL PROTECTED]                                     Subject 
             .EDU>                     Re: GRS / ICF                       
                                                                           
                                                                           
             12/04/2007 06:06                                              
             AM                                                            
                                                                           
                                                                           
             Please respond to                                             
               IBM Mainframe                                               
              Discussion List                                              
             <[EMAIL PROTECTED]                                             
                   .EDU>                                                   
                                                                           
                                                                           




Hi, Matthew.

Yes,  I want to create a Coupling Facility inside  a single machine.
Previously I have had GRS in ring mode,  but there were problems of enqueue
in the CICS regions on one of the LPar.
So now, I'll look the ICP CHPID  definitions too.

Thanks so match.

Saludos.
Álvaro.

-----Mensaje original-----
De: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] En nombre
de
Matthew Stitt
Enviado el: Lunes, 03 de Diciembre de 2007 21:32
Para: IBM-MAIN@BAMA.UA.EDU
Asunto: Re: GRS / ICF

I assume you want to create a Coupling Facility in the machine also.  What
you want to do is very feasible.  I suggest you look at the ICP CHPID
definitions also.

I have this running with 8 MVS systems connected to a single CF inside a
single machine.  Works quite well.

And I do not want any "If I did it" arguments, this is the hardware
limitations I must work with.  <g>


On Mon, 3 Dec 2007 15:30:09 -0600, Alvaro Quintupray B.
<[EMAIL PROTECTED]> wrote:

>Is it possible to install a GRS between two LPAR of only one machine using
ICF?
>
>Thanks
>Alvaro.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to