View previous topic :: View next topic
|
Author |
Message |
Willem Vermeer
New User
Joined: 31 Oct 2007 Posts: 38 Location: Amsterdam, the Netherlands
|
|
|
|
Greetings.
First, something on our environment. We're running z/OS V1R11 and z/OS V1R11.
Now our problem. I have a couple of productionjobs, run by people who do versionmanagement of all kinds of productionstuff (mostly programs). They first transfer data towards the system, that is to be updated, and allocate it there as a <dataset>.NEW.
Then, they run a job in which all the <dataset>s are allocated (in the JCL) with DISP=SHR (Significant? I don't know). This job then does the following:
1 - DELETE the <dataset>.OLD
2 - ALTER <dataset> NEWNAME(<dataset>.OLD)
3 - ALTER <dataset>.NEW NEWNAME(<dataset>)
The strange thing is, that now I find several data sets that [a] have been uncataloged (presumably by a DELETE) and that [b] DO NOT have a NVR in the volume's VVDS. We're dealing with SMS-managed volumes.
The first indication that something was wrong, was the fact that they ran into a DUPLICATE DATA SET-error because the volume still contained a <dataset>.OLD. So the first ALTER failed. No errormessages have appeared. No IDC3009I or something like that.
Anybody any ideas?
Willem Vermeer
ING Bank
the Netherlands |
|
Back to top |
|
|
Bill O'Boyle
CICS Moderator
Joined: 14 Jan 2008 Posts: 2501 Location: Atlanta, Georgia, USA
|
|
|
|
Willem,
I'm assuming this is a VSAM KSDS file.
In your ALTER are you issuing this against all three components, Base Cluster, Data and Index?
Mr. Bill |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10872 Location: italy
|
|
|
|
a quick and dirty reply ...
I remember that the rename of SMS managed dataset was a bit murky
when the old/new HIGH level qualifiers of the datasets were aliased to different user catalogs
but .... as I said a quick and dirty reminiscence |
|
Back to top |
|
|
Willem Vermeer
New User
Joined: 31 Oct 2007 Posts: 38 Location: Amsterdam, the Netherlands
|
|
|
|
No, as it happens these jobs don't do their stuff using KSDSes. It's all quite ordinary partitioned data sets (PDS). Not even PDSE.
Willem Vermeer
ING Bank
the Netherlands |
|
Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009 Posts: 580 Location: London
|
|
|
|
If the newname hlq alias points to a different usercatalog than the original hlq then the datasets will be renamed but will still be cataloged in the old catalog. They would not be visible through the normal catalog search order, you have to do a listcat with the catalog parameter pointing to the old original catalog. The NVR/VVR's would match the newname and would be visible on the volumes where the datasets reside, but if you printed the volumes VVDS you would see that these have a backward reference to the original hlq catalog. The newname entries could be mergcat'd across from the oldname catalog to the newname catalog and would then be visible normally.
This can be a danger of renaming. It is better to copy to the newname for that reason as the newname would be created and cataloged correctly. |
|
Back to top |
|
|
|