Blindingly fast disk I/O and problems with 1542 ans SCSI tape

Karl Denninger karl at ddsw1.MCS.COM
Sun Feb 18 12:59:33 AEST 1990


In article <1171 at digiw.UUCP> petri at digiw.UUCP (0000-Admin) writes:
>	I have used adaptec
>	1542 controllers, but ihave had big problems with Archive
>	150 M SCSI streamers. I have got always PANIC .. error code 17
>	when i have used steamer . I am now looking some other
>	solution for SCSI disc and tape.

>	The problem with 1542 and Archive 2150S is not problem with
>	only one unit , i have tested 5 diferent ones of 1540 / 1542
>	three 210S`s and in many diferent computers.    .... And
>	once i got it working, but then i sold controller ... !.
>	The problem is not with SCSI id , not with parity .

You've got problems with the SYSTEM; that drive/host adapter combination is
known good!

Check first that the system works ok with the adapter.  From MSDOS, go into 
DEBUG, and do:

	g=dc00:9

This will run a test for the first-party DMA.  If this test does not pass
for the entire address range, you have a problem with the motherboard you
are using and MUST fix it.  Otherwise you will eventually see nasty
results, including bad data on the disk, panics, and other untracable
problems.  Note that this test is not documented in the manual.

IF this test passes, you should be able to use the board without trouble.
Note that the test will take up to 5 minutes per segment, and it checks each
segment of memory separately.  Take the time to let the test run over the
entire base RAM.

IF THIS DOES NOT PASS try changing the Bus clock rate if you can (some 
motherboards can change I/O waits, clock speeds, etc).  This shouldn't
matter, but once in a while it does.  If you still can't get the board to
pass diagnostics DO NOT USE THE AHA154x WITH THAT MOTHERBOARD.  Doing so is
inviting some really nasty problems, none of which you will be able to
isolate.  We've seen bad disk writes (ouch!) and panics on boards which
don't pass the diagnostics.  

Note that ISC by default will ignore the sync jumper and DMA transfer rate
jumpers on the AHA board.  We have patches to enable these functions in
test; for now, pull the DMA rate jumpers and SYNC ENABLE jumper while you're
running the DMA test from the ROMs in order to simulate the actual rates
you'll be using under ISC.

We sell this combination of components, and use it in-house.  I've yet to 
have a problem with the Archive 2150S and AHA1542 in a motherboard that 
correctly supports the first-party DMA that it uses.  (We can also write
125MB tapes in the 2150S, although ISC specifically says that won't work...
you just have to know which minor devices to use :-)

--
Karl Denninger (karl at ddsw1.MCS.COM, <well-connected>!ddsw1!karl)
Public Access Data Line: [+1 708 566-8911], Voice: [+1 708 566-8910]
Macro Computer Solutions, Inc.		"Quality Solutions at a Fair Price"



More information about the Comp.unix.i386 mailing list