We have two IODF's for our entire shop, and run mulitple sysplexes (unfortunately). While we're looking toward merging systems, we will continue to have a test (sandbox) and PPRC watchdog systems that are logically separate from our production sysplex, but share gadgets (we do not have the funding to physically isolate our sandbox). We are GRS only, no MIM/MII/MIA.
For those that were interested, here's how we do it: 1. Separate buildings, one z9 in each, each with it's own DASD. Our watchdog systems have their own control units, accessible from all systems within it's building, and from the watchdog in the other building. 2. IODF's are separate HLQ's, aliased from all master catalogs to one common user catalog per building. 3. User catalogs are isolated on watchdog volumes, and excluded from backup on all systems except the watchdogs. No other data is on the catalog volumes. OS Configs specify OFFLINE everywhere except the owning system. 4. IODF's are are isloated on watchdog volumes, and excluded from backup on all systems except the watchdogs. No other data is on the IODF volumes. OS Configs specify OFFLINE everywhere except the owning system. 5. RNL is set to EXCL TYPE(GENERIC) QNAME(SYSIGGV2) for the IODF user catalogs. All other catalogs are CONverted. We currently EXCL TYPE (GENERIC) QNAME(SYSVTOC) becase we are still in transition from RING to STAR. Once in STAR, we will CON, *except* the IODF volumes (EXCL TYPE (SPECIFIC)). This allows the RESERVE to be issued against the IODF volumes to maintain integrity. 6. We use HCM. We connect to one of the two watchdog systems to perform all configuration updates. When we implement, we build production, write IOCDS, copy the IODF to the other site, write IOCDS, and ACTIVATE TEST on the watchdogs. If all is well, we HW/SW ACTIVATE on the watchdogs, then VARY the IODF volume ONLINE to the other systems in each site in turn, SW ACTIVATE, then VARY the IODF volume OFFLINE. Of course, if there are DELETE's or POR requirements, we handle those as appropriate. As we only have ONE person performing IODF updates, and we isolate the data, RESERVE is not a performance or availability hit. We know this is not sanctioned, but we manage it by single-threading and adhering tightly to documented process. Regards, Arthur Gutowski Ford Motor Company ITInfrastructure [EMAIL PROTECTED] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
