> On Sun, 22 Jun 1997, Gerard Roudier wrote:
>
[ ... ]
> hmm, how can i describe my scsi-bus ...
> i will go from one terminator to the other ...
> first end (inside computer)
> |--- Quantum LPS with terminators installed (passive ?) --->
> ---> IBM DCRS (no way to terminate this one ... <g>) ---->
> ---> NCR 810 --->
> ---> external cable (80cm) --->
> ---> external 4-device-box --->
> ---> Seagate ST1600N --->
> ---> Sanyo CD-ROM --->
> ---> HP 6020i --->
> ---> end of external box --->
> ---> external active terminator ---|
> makes a total of no more than 2 meters ...
Perhaps a little more, since if the external box has 2 connectors, the
internal cable length should be about 80 cms.
You SCSI system uses 10 connections and is loaded by 5 devices + the
controller.
Each connection and devices introduce electrical perturbations,
especially capacitances.
The Ultra drive has probably been tested only with Ultra SCSI compliant
SCSI stuff.
Looks like a
ShouldWorkInTheoryButMayHaveEveryChanceToGetProblemsScsiSystem (TM). :)
1 - Seems you are'nt sure internal terminators are active.
If the DCRS has terminators, they probably are active ones and
you could use them instead.
Are you sure all devices, including the controller, that shall not be
terminated are correctly configured with terminations disabled.
2 - You are using internal and external devices. Such configuration is
sensitive to cable quality.
> > We probably should expect such problems to be recovered, hewever,
> > error recovery is very hard to implement and to test and, in any case,
> > it is not possible in my opinion to recover from all kinds of errors.
> perhaps i should try to avoid the error ... :)
Exactly.
[ ... ]
> > > seems to happen only if the system is running under
> > > heavy load AND the ST is powered up some time ...
> > Do you mean that you powered up the ST while the system is running?
> yup. it is connected to the bus all the time, but normally
> powered down (no 5 and 12 V ...) .. and after the "booting"
> of the disk and an add-single-device everything worked for
> more than an hour. and then --->> KABOOM !
Forget 'add-single-device', unless you are sure that the device you
switch on/off while the SCSI lines are driven does behave correctly
in regard of SCSI lines when it is switched on/off.
Since you probably are'nt quite sure of that, you should only switch
boxes in that order, in my opinion:
1 - switch on the external box, then switch on the system box.
2 - switch off the system box, then switch off the external box.
> > > (can an overheated hdd data-kill another one via the scsi-bus ?)
> > Since the SCSI bus is a shared resource, any device on the bus can
> > make the resource unusable.
> i guess a scsi-bus-testing-device costs more than a new harddisk ? :)
I donnot understand the joke, unless my obvious answer was not clear.
An overheated hdd is more usable for frying eggs that to shared a SCSI
bus, if you see what I mean. :)
[ ... ]
> ST is switched in an "kind-of-idle" state ... but as i said
> before, it works for quite a long time after power-up !
What can I do ...
[ ... ]
> yup. my bet is the overheated ST ... i will try to
> find out, if i am right.
> > My recommendation is to use more than 1 scsi BUS and to distribute devices
> > among buses in a way that will minimize the risk to get SCSI problems.
> > 2 buses is generally enough for most systems.
> > Base choice on speed, purpose, age, quality, etc.. of scsi devices.
> > That cannot be bad, at least for performance when you are using 2 devices
> > with very different speed at the same time.
> smile, i would like to use a 3940, if you give me one for free ... :)
Using 2 NCR53C810 controllers offers also 2 SCSI buses for far less
money.
[ ... ]
> i think my bus is ok, just the drive drives it crazy ...
Seems that only a single response was acceptable for you.
Bad spot for me! I'm not been able to guess it. I will try to do better
next time. :)
Gerard.