Date: Mon, 7 Oct 96 15:07:35 CDT From: bartling@cyrix.com (Steve Bartling) To: bartling@mega.cyrix.com (Steve Bartling) Cc: freebsd-hackers@freebsd.org, freebsd-questions@freebsd.org, dietz@cyrix.com (Carl Dietz) Subject: Re: Irwin ( exabyte 8200 ) and Freebsd 2.1.5 RESOLVED Message-ID: <9610072007.AA19775@mega.cyrix.com> In-Reply-To: <9610031817.AA04092@mega.cyrix.com>; from "Steve Bartling" at Oct 3, 96 1:17 pm
next in thread | previous in thread | raw e-mail | index | archive | help
For those of you with the 1742A or other aging Adaptec products ... I spent 1 week trying to determine why an Exabyte 8mm tape drive would not work under Freebsd 2.1.5 or Windows95 with an Adaptec 1742A. Symptoms ... : Freebsd 2.1.5 would properly recognize the tape drive upon boot and could read tapes, but not write tapes. Attempting to write the tape would cause SCSI bus lockup ( a hard reset is required ). Windows95 could recognize the tape drive with various backup programs ( Arcada, etc ). However, attempts to write the drive would lock up the SCSI bus. Unlike Freebsd however, the drivers under windows 95 could recognize the bus problem and would reset the SCSI bus. I would really like to know how this is done. Solutions : The Adaptec 1742A bios/microcode was not up to date. My card had the following designations on the bios/microcode eproms : 450113-00 D MCODE B7D6 450216-00 A BIOS DC00 Downloading the latest bios/microcode from Adaptec's web page fixed the SCSI bus hang condition. There are two firmware updates on their web page. The firmware file named 174XA.EXE has a corrupted bios file. Do NOT use this firmware. The firmware file named 174XAS.EXE has the more recent bios/microcode and is not corrupted. Please make sure that you use the same EPROM types/speed as the original board uses. Please note that the bios revision number that is displayed at boot time does not change ( maybe microcode is the only thing changed ? ). I do not know what was wrong with the old bios/microcode as Adaptec's on line documentation on the firmware upgrades was nil/non-existent and I never heard back from their technical support. Also, the switch settings on the MX controller card for the Exabyte tape drive need to be tweaked. The card was set for fixed block size mode. The best setting is variable block mode. The parity checking was disabled ... this needs to be enabled if the rest of your scsi peripherals have parity checking enabled ( like mine ). Remember, parity must be ON for all devices, or OFF for all devices on the same scsi bus. Another useful thing to change is the memory check during scsi bus reset ( or power-on reset ). It takes about 60 - 70 seconds to complete this test and this will delay your DOS/windows95/unix boot until the test is completed. If you change this setting to disable this test, boot times are greatly enhanced. Note: reaching these switches is a pain in the butt. For the switch settings, please see the Exabyte web page at http://www.Exabyte.COM/TechnicalSupport/ProductSupport/answers/in0111.html I want to thank those of you who responded. I especially want to thank everyone at Exabyte for the following reasons : 1) Exabyte actually supports old products. Their web page is loaded with information about older out of production stuff. They even provide invaluable DOS based tape drive test programs. Every firmware revision is thoroughly documented so that the history is evident and you can evaluate if you need to upgrade your current firmware. Most tape drive vendors only have info on their current products in their web pages. Believe it or not, I actually decided to purchase an Exabyte 8mm drive instead of other vendors 4mm drives due to their support for old products. After spending 3 weeks browsing web sites, BBS sites, old product reviews, etc ... I decided on Exabyte. 2) Exabyte actually responded to my email about the problems I was having. This is in stark contrast to Seagate and Adaptec from whom I never heard anything. In addition, the support personnel at Exabyte are well trained and knowledgeable. Many thanks to Lance Blumberg at Exabyte !!!. Thanks to everyone who responded ... - Steve Bartling P.S. My original message is appended below. > > Howdy, > > According to the Freebsd 2.1.5 documents, the st scsi tape driver > has built in support for the Exabyte 8200. > > I have attempted to install this drive on SCSI ID 5. (Side note, > the firmware revision is the generic 2618.) > > I have used the "expert" DOS software package that I downloaded > from www.exabyte.com to test the drive under DOS. Other than > a tendancy to report fairly high Rewrt and ECC percentages on > the first couple of megabytes written, the drive passed all > of the tests that this software provides. The Rewrt and ECC > percentages will drop as you write/read more data from > the drive. If you write/read enough data, the percentages > average out to acceptable values. Is this typical behavior ? > > Oh well, I digress ... > > My real problem is with Freebsd 2.1.5. > > I have two tape drives installed. An Archive 2150s is present > on SCSI ID 4. The 8200 is installed on ID 5. > > I created the proper devices in /dev using the "/dev/MAKEDEV st1" > command. All of the proper devices were created. > > I rebooted just to make sure everything was clean and verified > that Freebsd properly probed/detected the 8200 and attached it > to the st1 interface. The Archive 2150s was also properly identified > and was attached to st0. > > Now for the problem ... > > The Archive 2150s works fine. When I attempt to "tar -cvf /dev/rst1" > the entire SCSI bus hangs, requiring a machine boot to clear the > bus. I cannot use the st or the mt utilities to display status or > set modes. When I attempt to do so, I get "Invalid Input/Output Error". > > Again, I can access the 2150s just fine. Hardware does not seem to be > at fault since both drives are happy under DOS. > > Any clues ? > > - Steve Bartling > > email: bartling@cyrix.com >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9610072007.AA19775>