On Thursday, 07/10/2008 at 01:44 EDT, Marcy Cortes 
<[EMAIL PROTECTED]> wrote:
> Just trying to think through the best way - the trade off is tests
> automagic (ipl away) with disaster test or longer involving reconfigging
> with real diaster (mucking with system netid perhaps).  Given the former
> happens 11 times a year and the latter hopefully never, maybe its ok for
> a little intervention in a real disaster scenario (given we meet our
> RTO).  But its also probably better to have an RTO in test that is the
> worst case scenario and not better than the real disaster.
> 
> Didn't tcpip used to work off of system identifiers and not node names
> (system netid)?
> That latter is harder to work around given a change to system netid
> involves resaving CMS.

TCPIP will look for
1. userid TCPIP
2. nodeid TCPIP (from IDENTIFY)
3. PROFILE TCPIP

The default nodeid from IDENTIFY *is* the system id.  To test:
1. Copy SYSTEM NETID to your A-disk.
2. Delete all lines in it, except a comment.
3. IPL CMS PARM NOSPROF
4. ACCESS (NOPROF  (optional)
5. IDENTIFY

Alan Altmark
z/VM Development
IBM Endicott

Reply via email to