On Thursday, 08/09/2007 at 03:48 EDT, "Schuh, Richard" <[EMAIL PROTECTED]> 
wrote:
> Here is one forAlan, but it looks like Chuckie might have a comment, 
too.
> 
> During our plight with TCPNJE, we tried routing the jobs through an 
> intermediate MVS system. The jobs reached their destination and ran. The 

> destination MVS system then tried to send the results back by the 
reverse 
> route, The output files reached the intermediate MVS system and RACF 
stepped up 
> to the plate. The files were allowed to go no farther because there is a 
userid 
> on the MVS system that matches the VM userid for which the files were 
tagged.I 
> know that this is MVS RACF, but maybe Alan has the answers to these 
questions:
> 
> 1.     Is that normal behavior for RACF? Would RACF on VM do the same 
thing?
> 
> 2.     Why is RACF on MVS even looking at the userid of files that are 
tagged 
> for another node? It is the destination node that should be looking 
after the 
> security for the files. RACF on MVS has no idea what rules are in effect 
on the 
> destination system. Bydefinition, it is remote and possibly using a 
different 
> ESM.

I think I'd open a PMR with JES as they are the ones that make SAF calls, 
invoking RACROUTE.  I can't speak for z/OS, but on z/VM, any control on 
the destination of a file is handled by the originating system.

As to why JES made a SAF call, it makes no sense.  If the destination_node 
<> this_node, then this_node shouldn't be looking at destination_user.

Alan Altmark
z/VM Development
IBM Endicott
  • RACF Schuh, Richard
    • Re: RACF Alan Altmark

Reply via email to