sheldon@quest.UUCP (Scott S. Bertilson) (02/21/89)
I started using Kennedy tape drives many years ago with the 9300 and grew to appreciate them immensely enough to recommend them to anyone who asked. About 2 years ago, they brought out an "improved" rack mount model called the 9401 and a self-loading model called the 9610. We have found that these drives have a very nasty tendency to return a corrected error on records that should actually be reported as hard errors in GCR mode. It has been masked by the fact that the controller we are using reports a short record error with higher priority than a correctable error. As a result we see tapes written on mainframe machines with constant block size come in with wildly variable record sizes and no errors. When we try to read the same tape on a Cipher drive, we get hard errors. This is simply the most obvious demonstration of a problem which is far too common even with good tapes. Seems awfully likely that they have a bogus ECC algorithm, but who knows. We have tried to pursue this with Kennedy through our local rep, but haven't found them too interested in looking at the problem. I'm wondering if anyone else has seen this problem and whether there is a solution available. -- Scott S. Bertilson ...uunet!bungia!quest!sheldon scott@poincare.geom.umn.edu