View previous topic :: View next topic
|
Author |
Message |
m dhiraj
New User
Joined: 22 Oct 2009 Posts: 6 Location: Chennai
|
|
|
|
Whats the diffrerence between going for a Full System Replacement and a Software Upgrade...when we do z/OS installation using Serverpac.
Can anyone elaborate on the use of Master Catalog for Software Upgrade.. Should we use the existing Master Catalog ? If so... can anyone stress upon the use of existing master catalog for z/OS osftware upgrade |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
Google is your friend. Googling serverpac full system installation returned 352,000 hits; one of them on the very first page says
Quote: |
Two types of ServerPac installation are available to you. (You choose the type when you install, not when you order.)
* A full system replacement installs a complete z/OS system. It installs all the data sets you need to IPL, to log on to the target system, and to run a z/OS image in order to complete other installation and customization tasks. The installed data sets fall into two major categories:
1. System software and related data sets (such as distribution and target libraries, SMP/E CSI data sets, and sample libraries)
2. Operational data sets (such as page data sets, system control files, and a master catalog).
Because IBM creates a working set of operational data sets for you, a full system replacement helps assure a successful first IPL.
Depending on your environment, you might need to merge your existing operational data sets with the data sets created by ServerPac. You can do this before or after first IPL.
* A software upgrade installs only system software and related data sets (such as distribution and target libraries, SMP/E CSI data sets, and sample libraries). It does not create the set of new operational data sets required to IPL (such as page data sets, system control files, and a master catalog). With a software upgrade, all operational data sets are assumed to already exist and to be usable by the new level of software installed. When new operational data sets are required, you must allocate and initialize them before you IPL. For example, you might need to add parameters required by the new software level or change data sets so they will work with both the old and new levels.
A software upgrade uses your existing catalog structure. This includes your existing master catalog (with direct or indirect cataloging references) and user catalogs. In addition, software upgrade allows you to create new user catalogs as part of the installation process.
A software upgrade is possible for z/OS but not for the subsystems. |
If your question is asking about using the existing master catalog for a serverpac software upgrade, the answer is in the quote -- the existing master catalog will be used, period.
If you're wanting to overlay the current system, use Software Upgrade. If you are defining a new LPAR, you would want to use full system installation. Which you use depends upon what you're attempting to do. |
|
Back to top |
|
|
m dhiraj
New User
Joined: 22 Oct 2009 Posts: 6 Location: Chennai
|
|
|
|
Thank you for your suggestion.. I have already gone through the lines you have posted...
We are following the direct cataloging...
We have used the driving system master catalog itself...for the target system( as driving system = target system in software upgrade)
When we do practically the z/OS SOftware Upgrade.. the first job we encounterd is...DEFSSA. In this job, first it deletes the existing Alias entries frm the master catalog and then try defining them in the later steps...But, as the datasets already exist in the driving system, alias would not be created. So, how to overcome this... if we use the driving system master catalog for target system also. |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
It's been a while, and I'll need to talk with one of my coworkers in a few hours when he's in, but as I recall we exported the master catalog aliases and imported them. But it has been a while so I'll need to double-check. |
|
Back to top |
|
|
m dhiraj
New User
Joined: 22 Oct 2009 Posts: 6 Location: Chennai
|
|
|
|
Thanks Robert...
We will be waiting for your reply.... |
|
Back to top |
|
|
PeD
Active User
Joined: 26 Nov 2005 Posts: 459 Location: Belgium
|
|
|
|
I am a bit surprised that such touchy operations can be led via a mainframe forum !!!
And I don't speak about or against the big knowledge of Robert.
Cheers
Pierre |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
Personally, Pierre, I wouldn't bet my LPAR on anything I found on a forum -- that's what the manuals are for. |
|
Back to top |
|
|
PeD
Active User
Joined: 26 Nov 2005 Posts: 459 Location: Belgium
|
|
|
|
.... and years of experience ( does it sound like old fashion ???? ) |
|
Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007 Posts: 8797 Location: Welsh Wales
|
|
|
|
PeD wrote: |
.... and years of experience ( does it sound like old fashion ???? ) |
No Pierre, Just good old common sense. |
|
Back to top |
|
|
nevilh
Active User
Joined: 01 Sep 2006 Posts: 262
|
|
|
|
Quote: |
In this job, first it deletes the existing Alias entries frm the master catalog and then try defining them in the later steps...But, as the datasets already exist in the driving system, alias would not be created. |
Sorry this does not make sense. If you have deleted them from the catalog you can define them later . The catalog does not care if the datasets physically exist or not. |
|
Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009 Posts: 580 Location: London
|
|
|
|
I hope you're not doing this on a Production system!
I'm not familiar with ServerPac process, but be aware that if you delete the Usercatalog ALIASes from the Mastercatalog then any newly created datasets that have a name matching those aliases will from that point be cataloged into the Mastercatalog. If that has happened you won't be able to define the aliases in the Mastercatalog until you delete or uncatalog the actual files names from the Mastercatalog. Also, any datasets matching those aliases that were previously created, before deleting the aliases, will not be visible without specifying whatever Usercatalog they're were in.
Mastercatalogs should only have base system type data in them as a rule, SYS1.** and linklists etc. For anything else you should have a Usercatalog ALIAS (which can be from 1 to 4 levels of the dsnames, best to stick with 1) which points to a Usercatalog where they will be cataloged. After that any dataset create or search for a dataset matching that alias will be directed to the related Usercatalog. |
|
Back to top |
|
|
|