I use a similar setup in the boxes I sell, DC390U2B, IBM DMVS09V.
The harddisk is without a terminator, and I strongly believe yours
don't have it either. Are you sure about the DMVS36G, shouldn't
it be a D at the end. To get it working, a active terminator specially
made for LVD use must be at the end of cable. The controller has its
own internal automatic termination. (Yes, you can run the setup in SE
mode also, but that's a waste).
Please check that first before you dwelve inte further debuggin.
One thing you could do is to enter the scsi bios (F2 at boot) and
verify the disk (it is nondestructive, verify that the scan start
at block 0).
/Karl
-----------------------------------------------------------------------
Karl Hammar Asp� Data [EMAIL PROTECTED]
Lilla Asp� 2340 +46 173 140 57
S-742 94 �sthammar Unix for the small company
Sweden +46 70 511 97 84 (mobile)
-----------------------------------------------------------------------
From: "Kai Harrekilde-Petersen" <[EMAIL PROTECTED]>
Subject: Timeouts on ncr53c895
Date: Wed, 15 Mar 2000 15:32:27 +0100
> Hi guys,
>
> I know this is the kind of problem reports that everyone hates, but I
> would appreciate comments and suggestions from more experienced people.
>
> The setting: I am setting up a file/computation server here at work.
>
> The box is an Athlon-700MHz, 768MB RAM (append mem=758M set in lilo.conf),
> with a Tekram DC-390U2W LCD controller (ncr53c895 based)
>
> The SCSI channel has only one device; a 36Gig IBM DMVS36G (Rev 0100)
> disk. The ubiquitous(sp?) CD-ROM is on ATAPI.
>
> The IBM disk is a SCA disk, with a LVD converter kit on. I haven't
> checked the termination (yet).
>
> After installing Debian 'slink' 2.1r4 on the box (and compiling
> a 2.2.14 kernel), I tried to run 'bonnie' on the disk, just for kicks.
>
> I set default tag queue depth = 8, max tag queue depth = 64,
> and Max Sync frequency to 40MHz during kernel configuration.
>
> During the getc() test (char-by-char read), the scsi subsystem spewed
> 7 timeouts at the console, and the computer locked up (hard. I had to press
> the reset button to get it back to life).
>
> The timeouts said "scsi0, channel0 id0, lun0 Write (10) 00 01 1c db d6 00 00 02 00"
> (copied from the screen by hand). There were a some serialno's, which I didn't write
> down (I expect they are of little interest).
>
> The following six timeout were identical except that the "d6" were incremented by
> two each time (d6, d8, da, .. e2).
>
> I certainly do not expect anyone to try to debug the problem from these vague
> details, but I'd be very interested in hearing if anyone had bad (or good)
> experiences with a similar setup.
>
> Since I saw signal 11's during kernel compilation, I'm inclined to suspect memory
> timing problems, but I prefer to keep an open mind until I am able to catch
> the villain red-handed for sure.
>
> Thanks for any comments,
>
>
> Kai
> --
> Kai Harrekilde-Petersen
> Exbit Technology A/S
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to [EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
