gotwols@warper.jhuapl.edu (Bruce Gotwols) (03/31/91)
I recently purchased an Exabyte 8500 8 mm tape drive. (This is a new product with about 5 GB of storage. I am not talking about the older model 8200 here.) In the past I had successfully connected our older Exabyte 8200 tape drives to a VAX 3100 workstation by connecting it directly on the SCSI bus. However, I have had no luck whatsoever getting the new 8500 to operate on our VAX 3100 workstations. The problem is not a grossly malfunctioning Exabyte as I have succsessfully used this same drive on an IBM PC, a SUN 3 and a SUN 4. The symptoms of failure on the 3100 workstation are many. Beside the error log entries for the tape, I found that other disks on the same SCSI bus were experiencing errors while the tape drive was physically connected. I also tried writing to the tape and then reading the data back in. Numerous errors (>50%) were found in the very first record read. In other words, the troubles are quite severe, and not just an ocassional glitch. Calls to the supplier have lead to two stories, neither of which I have yet been able to confirm from other sources. In the least believable story it is stated that DEC has made the SCSI bus on their workstations "slightly incompatible" with standard SCSI by switching two pins around. That seems unlikely since the older Exabyte worked fine on the 3100. The other story is that there is a timing problem in DEC's SCSI driver support for the 3100 workstation. I can't confirm or deny this story. I will note however that the new Exabyte drive is capable of running about twice as fast as the older model, hence timing problems are at least plausible. For the record I should mention that our VAX 3100 workstation is in a cluster running VMS v5.3. The Exabyte model 8500 is brand new and has not had it's ROM upgrade that will allow it to read tapes made by a model 8200 drive. If anyone has had similar problems or knows something about this I would appreciate hearing from them. If a number of replies are received I will post a summary. Thanks, Bruce -- Bruce L. Gotwols Johns Hopkins University, Applied Physics Lab., Laurel MD 20723 Internet: gotwols@warper.jhuapl.edu (128.244.176.48) SPAN: APLSP::STR::GOTWOLS