Yep, I've shot my self in the foot with CLASS A sufficiently enough in the '80s, that I don't have feet left.
Since that time, I've trained myself that when ever I'm doing a "bad" command, to verify, twice, where I am and what I'm doing. That's why class a is OK for me. Little things like I very seldom use the "force" command. I logon to the guest, verify it is the one I want to knock off, and then issue a logoff command. Same for shutdown (which system am I on?). And my biggie is formatting drives. After doing the diskmap and verify the disk or minidisk I want, For minidisks: Link the drive as 800 q v 800 (to see if it is the right size) q link 800 (to see if someone else has it linked) access 800 z (see if it is a labeled drive and the right one) then I do the format (and I only format vaddr=800 drives. For full drives. attach the drive as 800 q v 800 (to see if it is the right size) access 800 z (to see if it is already labeled) ickdsf fn ft fm (again I only init/format vaddr=800 drives) And to keep things even more confusing: I use 520RES for LPAR 0. 521RES for LPAR 1 522RES for LPAR 2 For the rest of the volumes: VM* is for LPAR 0 (old convention) L1cua is for LPAR 1 L2cua is for LPAR 2 Makes it easy to vary and attach volumes in the system config file. Tom Duerbusch THD Consulting >>> David Boyes <[EMAIL PROTECTED]> 8/22/2007 3:12 PM >>> > USER VMTEST xxxxxx 64M 64M ABG Eek! 1st level class A privs for a test id? That'll last as long as it takes for you to type CP SHUTDOWN to the wrong level of CP...8-) BG or BCEG should be more than sufficient, and there are very good arguments to make test ids like this class G only and do anything privileged (ATTACH, DETACH, etc) from another id or DEDICATEs in the CP directory entry. -- db