Oh ... I always forget something ...
Some may find it handy to leverage 1st level CMS disks.
(You can always change this later and let 2nd level have its own
CMS support disks.) For example, MAINT might have
MDISK 190 MNT190 0 END RR ALL
-or-
MDISK 190 DEVNO 190 RR ALL
and so forth for 19E, 19D, 319, 31A, and others. Same trick for TCPIP.
The (1st level) virtual machine may be running a different release
of CMS than your production VM host workload. No problem! In my
prior life, we kept old releases of CMS back to ... shoot, probably 10.
To top this off, have OPERATOR do an 'IND USER' and look at the
second line of that response. If it was IPLed from device, (try to)
reIPL via NSS. If that fails, redo the DEFSYS and SAVESYS steps.
That way you automagically get an IPLable CMS NSS and yet have
no maintenance to get it instantiated. Combine this with the CMS recomp
disk for booting guest CP, and you can safely blow away the volume
with your DRCT and spool (and warm start and checkpoint) and it all
gets re-generated nicely. Reproducability is a good thing.
-- R;
