And of course HSM duplexing for ML2 and Backup tapes automatically covers the requirement to use different drives. Since the two copies are made concurrently, that guarantees they have to be on different drives.
It should be obvious enough to not need saying, but you also don't make two copies of critical tape data by running a copy step to make a copy of the first tape: a drive failure or some physical accident to the media during the copy process may destroy the one and only copy of the data before the copy process completes. The original DASD-based data should be used to generate all tape copies. With modern tape technologies, actual failures due to the tape media itself tend to be gradual and correctable: during writing, data blocks are verified as they are written and marginal spots on the media skipped, and error-correction encoding provides redundancy and correction when reading data if surface damage occurs later. In a properly maintained tape library where cartidges that start to show issues from age and use are replaced, data loss from a media issue should be exceedingly rare to non-existent. That said, catastrophic tape failure can always be induced by a device failure that causes physical damage to the media; or by some accident, mis-handling of a cartridge, or environment disaster that results in physical damage to the cartridge & media. Logical failures are also always possible, where a series of operational mis-steps results in a valuable tape being marked for deletion and re-used before its time. Joel C Ewing On 7/9/20 7:49 AM, R.S. wrote: > Regarding tapes: this is one of the advantages of using HSM and > physical tapes. It's quite easy to manage that even very old backup is > on quite recently recorded tape. Migration from older tape system to > new one is piece of cake. For VTS things are a bit more complex but > still it is possible. "Fresh" tapes are better than 15-years old cart, > last mounted 10 years ago. Not to mention weared drives. > > And of course always use two physical tapes. Preferrably in two > locations. Even the best tape may fail. Two cart also may fail, but it > is less likely. Three tapes (in three ATLs, in three locations) are > even more unlikely to fail concurrently, etc. And it is your decision > to say "n copies is enough safety for me". And I'm sorry, I don't > believe in any support from business side. > > BTW: two tapes, but avoid to write or read them in same drive. Drive > failure may somehow destroy the tape. I know a guy who tried to > recover data from backup, but the tape was faulty. He had two copies. > Second reel was also faulty. Actually both tapes were mounted in same > faulty drive which destroyed both copies. In that scenario even dozen > of copies would not help (assuming still same drive). > -- Joel C. Ewing ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
