--- snip --- <----- snip -----> >these datasets will be PPRC "mirrored"
One of the two sets of datasets IBM recommends to not mirror is the CDS set. The other is PAGE. I have another set that I never mirror'd: SCRATCH. The overhead of synchronous mirroring kills the first two, for sure. <----- snip -----> The GDPS hyperswap is impossible without mirroring the PAGE datasets. We don't mirror couple datasets (we are GDPS environment), but mirroring the LOGR couple datasets is a must if you want to have the information about the logger staging datasets. As I understand it, the reason not to mirror the couple datasets (XCF couple dataset in the first place) is the possibility of freeze on the primaries. --- snip --- When your I/O load gets swapped to your alternate site, you must assume that you will not have *any* access to the primary site anymore (I'm only talking about the I/O infrastructure). If you want JOBs to continue uninterrupted because of the swap, then you need all the data in your alternate site. This includes all paging and also 'scratch' volumes. The couple datasets (volumes) are excluded from this SWAP process, since XCF can manage primary/alternate couple datasets quite well on its own (with no I/O pause needed to swap UCBs). That is why you spread your primary/alternate couple datasets (dedicated volumes) between both sites. John ---------------------------------------------------------------------- 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
