On Sun, Dec 15, 2013 at 4:24 AM, venkat kulkarni <[email protected]
> wrote:

> Hello,
>           We have 3 system in  plex and out of three systems, two are
> running with z/OS 1.11 and one with z/OS 1.13. commands ( whoami ,id )
> running on z/OS 1.11 running fine but on z/OS 1.13 system OMVS, these
> command giving bad output.
>
>
<snip>

Based just on the above, is the z/OS 1.13 system in the same GRSplex as the
other systems? Why?
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ichza2c0/4.9
<quote>
If your shared RACF database is at application identity mapping
(*AIM*)level 1 or higher, all systems that update the OMVS segment of
USER or
GROUP profiles, or update the ALIAS segment of general resource profiles
(for example, any SERVAUTH class profile), or run RACF utilities, should
have global resource serialization connections between the systems, should
be in the same global resource serialization complex, and should be running
OS/390 release 10 or any z/OS release. Adding or deleting a profile that
has any of these segments, altering these segments, or running RACF
utilities from a system outside the global resource serialization complex
might result in incorrect results; for example, an alias index entry for an
OMVS UID or SERVAUTH alias might point to the wrong profile, or to one that
does not exist. To prevent database sharing errors, it might be useful to
use RACF program control to restrict access to all RACF commands that can
update these segments, to ensure that they cannot be used from systems
outside a single global resource serialization complex.
</quote>

Also, I found a "curious" phrase here, towards the bottom of the page :
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ichza7c0/17.2.1
<quote>
For installations at *AIM* stage 2 or higher, you can list a set of users
or groups with a specific UID or GID, for example using '223' for the UID
value and '55' for the GID value, enter:

         SEARCH CLASS(USER) UID(223)
         SEARCH CLASS(GROUP) GID(55)

</quote>

This is in the z/OS 1.13 manual. And it says that you can do a SEARCH for a
UID or GID at AIM level 2 or higher. But does not mention exactly what
happens at AIM 1 (your level) or lower. I wonder it the z/OS UNIX command
you are using has a similar restriction. I don't know. I'm still looking.

Looking in the UNIX manuals, I would _guess_ that these commands use the
UNIX program, BPX1GPU, to get the information. This, in turn, uses the RACF
callable service IRRSUM00 (aka GetUMAP). If this has some problem, it
writes a LOGREC record, per
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ichzb2c0/2.4.1 .
If worse comes to worse, you might want to dump and print an EREP report
quickly after a failure to see if RACF has written any error information.


I can't really figure out exactly what the problem is. But I will say that
it would be __extremely advantageous__ for you to see about upgrading your
RACF database from AIM 1 to AIM 3. Most of the new facilities and "speed
ups", especially in UNIX, require an AIM level of 3. It was very easy for
me. But that's because I had a very well controlled UNIX environment and
few users when I went from AIM 0 to AIM 3. It only took me about 3 weeks,
and that was only because I wanted to got to AIM level "n", then wait a
week, before going to AIM level "n+1". I have no problem occur due to this.

-- 
This is clearly another case of too many mad scientists, and not enough
hunchbacks.

Maranatha! <><
John McKown

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

Reply via email to