We run a similar cloning system with full SMP support behind each SYSRES. sysres 0 is the target of all new maintenance. the other two sysres 1 and 2 alternate roles from being the 'next' and being the 'minus one' versions of maintenance. Sysres 0 is used to shake down new maint. Then it is copied to either sysres 1 or 2 for permanent promotion.
I have no idea if the is THE best practice, but it is OUR best practice. There have been very few times when we would apply maintenance to a running sysres to fix an emergency, but the capability exists. I have noticed that every RSU level has included PTFs that are in ERROR. In this new world of Continous Delivery, every maint wave is a brave new world. This also increases the demand to put fixes in more often than in the good old days. We have settled on quarterly as a goal. Delaying and then being faced with over a thousand PTFs frightens me a bit. By using RSUs as guideposts we spend less time with problems that occure when trying to put on every PUT and every PTF as they are released. Every cycle starts with an ACCEPT. This gives a RESTORE/REJECT options until we decide we are ready to do another wave. I have never been a member of the 'Never ACCEPT' club. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
