I agree with the assessment that the incorrect policy was a ticking time bomb. In general, I highly recommend always using a new policy name for any changes no matter how trivial. 'New' in this case does not mean never used before, simply different from the name currently in use.
Decades ago we settled on a practice of using three names in rotation: 1, 2, and 3. Each policy replaces the previous one, with 1 eventually replacing 3, ad infinitum. While (re)using the same name is valid, you have no easy way to distinguish the 'new 0' from the 'old 0'. With just three names in rotation, differences can be spotted by inspection. If a policy is delayed pending some cleanup, you can see by the name alone that something is wrong. We also back up each policy source as a step in moving forward, so we have several 'generations' to look at when researching changes. Another recommendation is not to *change* the serial number in a policy when switching hardware but to *add* the new serial number. XCF will find only one connected, but the transition forward is much smoother. Even more so the case of fallback, where the new CF cannot be used for some reason. Once the new CF is stable and fallback no longer an issue, the old serial can be removed in POLICY+1. . . J.O.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 323-715-0595 Mobile 626-543-6132 Office ⇐=== NEW [email protected] -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Carmen Vitullo Sent: Monday, July 24, 2017 11:10 AM To: [email protected] Subject: (External):Re: What casues IPL/XCF to read the CFRM data set for the policy ? I generally do the same, but I do keep a backup policy defined just never activated, so really the only other way I think this could happen is if a policy change was pending for some reason, like a structure move from one CF to another, once the policy changes were no longer pending the old or invalid policy became active.? so you say "soon after the system were up and verified, an incorrect policy0 was written into the CFRM data set, had the wrong serial number and wrong partition number for the CF LPAR." you can define policies all day with no issues until they become started using a SETXCF command, or the PLEX is re initialized I believe. even if the incorrect formatted couple datasets I 'don't think' this would cause your issue, without testing this theory I cannot be 100%. ----- Original Message ----- From: "J Ellis" <[email protected]> To: [email protected] Sent: Monday, July 24, 2017 12:55:35 PM Subject: Re: What casues IPL/XCF to read the CFRM data set for the policy ? the name of the policy never changed, either in the parmlib member or the CFRM data set, the policy wa simply overlaid with the same name, but bad serial number and partition number. and, we ran that way for months ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
