Can't find drives on my dual channel Symbious SCSI card
 

Last year I bought a Kouwell PCI to Dual Channel Ultra SCSI card to handle multiple SCSI devices in my "experimental" Micron 400MHz Pentium machine that gets a lot of "castoffs."

Now, the SCSI devices I have "hung" on this machine include a UMAX Mirage IIse scanner and Nikon COOLSCAN III slide and filmstrip scanner. Currently, these external devices are daisy chained off an old Adaptec SCSI controller AHA-7850 that uses a 36 pin cable.

Internally, on another Adaptec SCSI card (AIC-7881U) I have the boot HD (ID-0), another SCSI HD (ID-1) on the 68 pin side of the adapter and a CDROM (ID-6) on the 50-pin side of the adapter. The adapter address is 7.

Then, finally, I have the "troublesome" dual channel card with an IOMEGA SCSI zip drive on one channel mounted internally at ID 6, and an external Netframe box containing four 9.1G drives at IDs 8-11 on the other channel.

All cards and drives are recognized and available when booted to Windows 2000 Professional :mad: .

When I boot to linux, either SuSE 8.0, or as of yesterday, Mandrake 9.1, I see and can use all of the SCSI peripherals except all four external drives on the second channel of the Ultra Card.

Now, when I started with SuSE 8.0, I had the "old" Symbios linux driver: sym53c8xx.o, but somewhere along the line, but certainly with the "upgrade" to Mandrake 9.1, I installed the newer Symbios driver with the "_2" extension to, at least, the *.gz file it was packed in.

Using SuSE's and Mandrake's hardware identifier, under the SCSI subsection, I get two lines of something like 53c1010-66 Ultra 3 SCSI adapter, and as I said before, the hardware identifier also "sees" the ZIP drive that is one one of the channels. It's just that when I look at the hard drive section of the hardware listings I can't see the four "external" hard drives in the Netframe box.

cdrecord -scanbus behaves similarly, I see sections for SCSI-0, SCSI-1 and SCSI-3, but nothing for what COULD be SCSI-2. When I go down /proc/, again, it's the same missing links to the devices.

Now the whole point of this exercise was to reduce the number of SCSI controllers, but until I solve this problem I'm getting nowhere. I like the flexibility of the SCSI bios over Adaptec's, and eventually, someone will have a faster SCSI device to "throw away" so I see myself "moving up," but not until I solve this precognition problem. There are so many places to look, the combinations of trials are endless.

So what I would appreciate is some suggestions as to what things, that I haven't tried to try. Now one area that slightly troubles me, but which I find very confusing is the discovery of then proper termination of LVD type drives. I think I got a "handle" on it this past week with the two internal SCSI drives. They had worked OK, but seemed system performance seemed particularly slow on the Windows side. I think I now understand the need (now perhaps more often than before) to expect to provide active termination on the cable rather than "expect" it on the drive. But then, I didn't see the system performance deterioration under linux on that bus and drives. And if I can use the Netframe drives under Windows, then why would linux have "balked" at them if it was a termination problem. I haven't "torn apart" the Netframe box yet, but I know it was a "hot swappable" raid box originally, so it should be configured however a 68-pin cable interfaces to the 100-pin hot swappable container, back to 68-pins, and I haven't read any hardware sites that explain termination of "hot swappable" drives.

But I don't want to influence a responses with the explanation in the last paragraph. It might not be a termination problem at all. If nothing else, I'd like to know if there are some commands that let me probe the card better. The kernel configuration file says that scsi_adapters are handled as modules.