> As I indicated at SHARE, for those of you who attended the wrap-up > session, my inclination is to: > > - create a second level VM, > > - apply maintenance, > > - see if it works, > > - if it does, do it all again on first level. > > > > We are a VERY small shop with only a handful of linux instances. We only > have z/vm in order to run linux. > > Can I put maintenance on with those instances up? Seems like that would > not be a good idea (comes to mind the time I did PUT2PROD from my desk, > TCP/IP got taken down, and I got zapped!). However, as Jim indicated, > z/vm 5.3 handles tcp/ip differently so ??? I would really appreciate > any insight that anyone can provide...
I don't put maintenance on with things up. I wait for a maintenance window or I don't do it. It is my gun and my foot, and I only have so many toes left. My procedure is pretty much the same as yours: 1) create second level vm 2) apply maintenance 3) test 4) wait for maintenance window 5) backup system 6) apply maintenance 7) test If I am feeling lazy and/or overly confident, I start at step 4. Which invariably causes a step 8) of _restore_ backups from step 5) and triggers a restart of the process at step 1) to figure out what went wrong.
