On 06/23/2010 02:56 AM, R.S. wrote: > Schwarz, Barry A pisze: >> Doesn't anyone exercise the DR plan anymore? If so, without reading >> the backups? > > Of course we do! Does anyone use PTAM for DR? Really? > We use remoted copy for DR. We don't need to recover anything from tape, > because we have all the data on DASD. Oh, we do have tapes with MIGRATs > and backups in DR site - we duplicate all the data on tape, and one > volume set resides in remote ATL. Without virtual tapes. > > BTW: Tape is fine for recovery single datasets. Time to recover all the > data from tape is simply unaccpeptable from business point of view. Not > to mention RPO for tapes (Recovery Point Objective).
Yes, really! PTAM to a remote secure vault, for transfer to appropriate DR site in event of a disaster. We would never consider using mass tape restores with only DFSMShsm or DB2 Utility facilities, as that WOULD take days. Generating full-volume DR dumps with FlashCopy for later staging to 3590-H carts involves a system disruption of only a few seconds, and restoring all the 100's of DASD volumes at DR from 3590-H drives with DASD and tape under FICON channels takes about 2 hours - and we then get to do some administrative tasks and twiddle our thumbs waiting on availability of the non-mainframe servers. It all depends on the nature of your business, whether you have legal requirements for sub-hour recovery times, whether recovery to a point-in-time with loss of several hours data is acceptable, whether the business can manage to run and survive for short-term, maybe even a day or two, without DP support, and what the risk factors are for your location. A management that pays for a DR capability that exceeds their requirements is simply wasting money, and companies the regularly waste money won't need DR after they go out of business. It would appear your company must have the luxury (and expense) of owning their own fixed-location hot site. I trust it is far enough from your primary site to not be exposed to the same failures. In the U.S. I've always felt that for unhardened centers a 30 - 50 mile separation should be minimum, but if you are anywhere near "tornado alley", in coastal hurricane States, in States subject to major storms that damage power distribution systems or cause flooding, or near known earthquake fault lines, that separation minimum should be several hundred miles. In the U.S. it is possible to contract for DR services at an un-owned hot site and maintain remote DASD and remote virtual tape there, but this is not cheap. Worse, the typical DR hotsite contract in the U.S. allows for the possibility in the event of a regional disaster that others who "declare a disaster" earlier may fill all the space in your normal hot site and you may end up in an alternate hot site separated from your remote data and have to copy all your data to an alternate site before running production. -- Joel C. Ewing, Fort Smith, AR jremoveccapsew...@acm.org ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
