> I therefore created a mini VM system with TCPIP started so that I could > restore it using this method and have an VM system that I could IPL. > > I did that a few years back. I'd can't remember off hand how that was > done. > Does anyone else remember?
This is a good thing to do. (in fact, I'll be giving a session at WAVV and probably the next zExpo on it -- another good reason to support WAVV...8-)) You need your CP nucleus and parm area, your IODF (if you use it instead of the much superior CP dynamic detection of devices...), your MAINT 190, most of TCPMAINT, and TCPIP itself. You also need a minimal OPERATOR, and I create one USERxxxx userid for each real tape drive you have (to allow parallel DDRs to occur on all your real drives). Configure USERxxxx to IPL 190 rather than CMS, so you don't need NSSes or spool space, and the USERxxxx ids don't even need a writable 191 -- you're not going to put anything on them anyway. FTPSERVE would be handy, but is not critical. On our 1-pack system, the PROFILE EXEC on the USERxxxx R/O 191 gets the userid, parses the xxxx and attaches the device at address xxxx as 181. The shared 191 also contains a DDR control file that restores from 181 to whatever is attached at 200 and some other assorted helpful tools (like TRACK, so you can be nosy and look at how each of the IDs is going. You don't need much (if any) page space, no dump space (unless you intend to use the 1 pack system for diagnostics too, for which you need spool and/or dump space); this system is supposed to be small and temporary -- just enough to get your real system restored as fast as possible. All those ids fit easily on one mod 3 sized volume (with a little bit of juggling and size management to remove some extra whitespace. One major caution: be sure the volume you're doing this on is in the Offline_at_IPL in your production SYSTEM CONFIG; you don't want this directory "accidentally" coming online if something happens to the production volume that contains your normal CP directory. I'm still working on the handouts for the talk; I'll post a draft when I get them finished. -- db
