The tape unit I use apparently has to be mounted on an IDE drive. I don't know if this is legitimate but that is what the builder told me. Apparently when it was physically installed it was attached to an IDE drive (8 IDE channels on the machine) and was defined to be ht0. I originally tried /dev/st0 which you mentioned and which did not work as it was not attached to a scsi interface. There is no SCSI on this Linux machine but the kernel has provision for scsi. The manufactures' box (and it was ordered with that intent) indicates Model STT28000A-RF ATAPI DRIVE by Seagate and also on the box was TapeStor Travan 8 GB. This indicates to me IDE mounted and that is what the builder intended. When I received the unit and before it was serviced (if I remember correctly) I could write and read from the tape using tar commands although there was always a read error but I'm sure I remember being able to examine the tape contents and extracting to a specific location. The read error was caused by a improper EOF which I traced to a corrupt MP3 file. But after the unit was serviced it was simply inaccessible. It would not even boot. Obviously I think the service company somehow messed up something controlling the Tape unit operation or the unit is itself faulty. After getting it to boot I think tar was writing to the tape unit but I am not sure. The tape unit starts whirring away but does not appear to work. So that tells me that maybe the Linux system at least knows where it is. The service company said they tested the tape drive and it was A okay but I'm not sure they tested it - after all they returned it all to me dead.
I can't take the Tape drive from the Linux machine and try in a win95 machine because I have no spare IDE channels on any win95 machines. I disabled the IDE raid and just went with conventional ATA66 drive operation tuned to maximum parallel operation to come as close to scsi performance as I could with 8 IDE channels. Anyway, prior to servicing the tape appeared to work off an IDE channel. At boot time there is this maze of errors as Linux attempts to mount the tape unit and access it, which it is no longer able to do, which sounds like a contradiction to Linux seeming to find it when I generate a tar command to /dev/ht0. This has nothing to do with the IDE Raid operation which has been disabled. I suspect this service company messed up the tape operation as they most certainly and without any doubt messed up the raid operation. They may have adjusted some reference files but do not appear to have messed up the operating system. It turns out they messed up the RAID by messing up the bios chips involved, this information I got from a consultant who worked on the unit after this service company returned the dead unit to me. It makes sense to me if what they did to the bios chips messed up the IDE RAID then what they did probably has something to do with the way Linux is now acting with regard to "ht0" the tape drive device identification. I am aware of the factors affecting this machine that could have been a problem and have ruled them all out. The only several possibilities that remain are the bios programming, the tape drive itself, changes made to reference tables, or possibly the need for special drivers as I do not know how Linux deals with IDE tape drives. If compilation were an issue here then the tape unit would not have operated in the first place. I have taken the time to explain all this so that you can appreciate the character of the machine with respect to the IDE channels and Tape unit and the reason why I am reluctant to recompile and specially so as SCSI operation because I do not believe it would work as a SCSI drive as you have been mentioning. It was a suprise to me that a tape unit was even built to work off an IDE drive but that appears to be the situation here. So if there are any other suggestions I will be glad to entertain them but I really think this tape unit is designed to be attached to an IDE channel. I have never had a Linux machine that involved multiple IDE channels as an extension from the two existing channels on the CPU board. I don't even know how to interroate the 3 bios chips that are involved. I can get at the chip on the main CPU but after that I am lost. It is quite possible, since this service company had altered these bios units, that the tape drive is no longer properly detected by one of these bios chips. So I think we should be looking at this TAPE unit problem from the perspective of what is required to make it work as an IDE unit. If I had recompiled for SCSI then I think I would have made a major mess. IF I can get the following help maybe I can get past this problem. 1. Is there a way of checking out the Tape, diagnostics or something that is operating system independent. Or do I just send it back to the manufacturer. 2. What information should be in the bios regarding an IDE based TAPE drive. 3. How can I access all 3 bios chips and what should I be looking for. 4. Are drivers required for IDE mounted TAPE drives and if so how to I determine where they are located, enable them, and that sort of thing. I think that any special compilation considerations were probably made when the unit was initially put together for me but what would I be looking for, where, and why. This whole thing is a mystery to me, apparently was also a mystery to the service company that messed it up, and apparently the consultant is not sure of what to do with the tape unit. So I am stuck with this expensive non working tape unit and there is probably just some minor little thing that is required to get it working. My bet is a bios setting somewhere or /etc/I_am_an_IDE_tape_unit_reference_file. ANY HELP WOULD BE APPRECIATED. Bye-thanks_TED Linux wrote: > > I have always used the Tandberg SLR series of drives for 5Gb to 100Gb. When > used with the Initio9100 UW controller they are detected and installed by > the OS and run perfectly without a hitch. Best of luck > > Mike > > -----Original Message----- > From: John Sanders [mailto:[EMAIL PROTECTED]] > Sent: Monday, 31 December 2001 04:48 > To: [EMAIL PROTECTED] > Subject: re: Tape Unit in Linux Box > > I use the Seagate Travan units. They are SCSI, so you should have SCSI > support in your kernel. The devices start at st0 (scsi tape zero). > Check to see if the device exists in /dev. > > I have 4 of these in one machine, so I have st0, st1, st2, st3 > > As far as I know, there is no "special drivers" for these tape units. > Tar should write to them just fine, however, I use an extremely old > version of bru to handle the backups. I believe bru just uses tar, > though. You might try a test with something like: > > "# tar -czvf /dev/st0 /some/temporary/non-critical/directory/*" > > then check the results with: > > "# tar -tzvf /dev/st0" > > Lose the quotes, of course. The -tzvf should give you a screen dump of > whats on the tape. > > Hope this helps. > > -John > > _______________________________________________ > Redhat-list mailing list > [EMAIL PROTECTED] > https://listman.redhat.com/mailman/listinfo/redhat-list > > This mail was processed by Mail essentials for Exchange/SMTP, > the email security & management gateway. Mail essentials adds > content checking, email encryption, anti spam, anti virus, > attachment compression, personalised auto responders, archiving > and more to your Microsoft Exchange Server or SMTP mail server. > For more information visit http://www.mailessentials.com > > _______________________________________________ > Redhat-list mailing list > [EMAIL PROTECTED] > https://listman.redhat.com/mailman/listinfo/redhat-list _______________________________________________ Redhat-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-list
