> From: sabell@ARGONET.CO.UK (Stuart Bell )
> Date: Sun, 11 Aug 96 21:56:53
> 
> Kjetil wrote:
>>>> I tried to use my Acorn SCSI card with my HP C3325A SCSI drive.
>>>> This drive was formatted 400 MB to RISC OS.
>>>> I am very persistant, and I this refuse to install it to an IDE drive
>>>> first as I did the first time I installed RiscBSD. As a matter of 
>>>> fact, I do not even have a suitable IDE drive for RiscBSD. The
>>>> 210 MB I have is my boot disc, and too small for RiscBSD anyway.
>  
> Stuart wrote:
>>> But you could put a boot partition on it, and put most other RiscBSD
>>> stuff on the SCSI drive? Surely the nature of beta releases is that we m=
> ay
>>> have to compromise a little on our ideal solutions, in order to get thin=
> gs
>>> running?
> 
> Kim wrote:
>> That is a totally unnecessary complication. I am quite sure there
>> are quite a lot of people that dont use RiscBSD because they
>> dont want the risk of repartition their harddisk, which probably
>> is full anyway. The second IDE slot is probably used up by a
>> CD-ROM.
>> The natural thing is to buy an SCSI disk for NetBSD, which use
>> SCSI disks on the other systems anyway. I bought an optical PD.
> 
> 1. If people don't want the 'risk' of re-partitioning their hard disc, then
> that implies inadequate back-up facilties in which case they shouldn't be
> contemplating RiscBSD. 

I have to agree, but still with good backup routines, some users
may still be reluctant to doing this.

> 2. The 'natural' solution of buying a SCSI interface plus a large SCSI disc=
>  is
> at least three times as expensive as buying a larger IDE drive. My net cost=
>  of
> going 420Mb to 1.6Gb was =A3135 inc VAT, and prices are falling. The SCSI
> equivalent, including buying a SCSI interface supported by RiscBSD would be
> well over =A3400.

Your point is that you can replace the 210/420/850 with a larger
drive. This is in fact a point.

But, this will not give you more performance. I doubt that this
will give you more than 2MB/s sustained transfer rate at the
most with optimized drivers.

To have more on the IDE drive, you will have to buy, e.g. a RapIDE
interface, and they cost approximately the same as high-performance
SCSI cards.

Also, the best IDE drives can not compete with the best SCSI
drives. My HP drive is capable of 4-5 MB/s sustained data
transfer rate, and this is slow compared to todays standard.
The best ones are approaching 10 MB/s (at least 7-8 MB/s).

At present, though the IDE drivers in RiscBSD are faster
than the SCSI ones, but this will improve with time.
It is just a matter of optimizing the SCSI drivers.

>> And what about me: I work professionally as a unix system administrator,
>> and know perfectly well how to develop unix programs, yet with NetBSD
>> on the RiscPC, I cant get anything done, and i have mounted on
>> a partitioned IDE drive. The reason? I dont have a HD on my
>> cumana scsi card. I have an optical PD instead. This makes the
>> boot sequence crash. Programming like that sucks.
> 
> No, it more likely means that the team can't be expected to write drivers t=
> hat
> will work with every possible drive or device, especially if they haven't g=
> ot
> access to those drives or devices.

You are missing the point. Kim had a 240 MB SCSI drive in his
Risc PC, but when he removed it, RiscBSD wouldn't boot. This is
what he is talking about. Not necessarily the fact that his PD may
not be fully supported.

>> However, I could rewrite it myself if I could boot and compile, which
>> I cant. The chicken and egg problem.
> 
> Back-up your Risc OS stuff to the optical; repartition to use almost the wh=
> ole
> IDE drive to boot and compile; write the RiscBSD drivers for the optical,an=
> d
> re-compile to get a system using the optical drive. Then repartition and
> reload your Risc OS stuff. Hardly elegant, but a solution to the 'chicken a=
> nd
> egg'. Unless your IDE is _really_ small. ;-)

This is completely off the point as Kim does not have problems with
disk space, but rather with getting RiscBSD up and running at all.

Kjetil B.