The chaining of volsers in the EOF record is where you have to ask the ta pe mount management subsystem for the new tape on a different address, get t he volser from its VOL1 record, write your EOF/EOVV records, detach that tap e drive, swap virtual addresses of the new drive to the old drive and then begin processing the new tape. It is indeed hairy with the RMSMASTR we ha ve now for VTS and ATL libraries. With a good Tape Mount management system, the process could be broken into pieces, where we ask the TapeMgr for a scrat ch volser, it mounts a scratch tape, verifies the label and returns us the DATA, then we process our EOF/EOV records, detach the tape and then ask t he TapeMgr to complete the scratch request by attaching that new tape to us. A bit more communications with the TapeMgr, but with the time it take to mo unt a 3590 tape in the ATL, perhaps that can be overlapped with our tape labe l gyrations.
/Tom Kern On Fri, 27 Oct 2006 12:33:34 +0200, Rob van der Heij <[EMAIL PROTECTED]> wrote: >The only hard part of it was multi-volume datasets. But the Piper made >me some enhancements to the tape stage to get that done (10-15 years >ago). I also have some recollection that the format required the >volser of the next tape to be in the EOF record, which made it pretty >hairy. >It's certainly doable, but I don't have access to tape drives anymore. > >Rob
