Steve - thank you very much for this detailed explanation. I have been
getting an education on SCSII drives more that I thought I would need but it
is seems to come in very handy in asking why things are not working the way
I understood they should.
In answer to Steve's question - Have you done the analysis in Table 10 in
the HASS manual?
This is what I have been using all along and yes and I had mapped with the
correct id in sysgen / ioconfig. I have found a cable with bent pins
(replaced that cable) and none of the cables were over 2 meters long and I
have had one drive that was bad - this has been replaced (does this sound
like Murphy's law at work) and my concern has been that the drives were
really DF and not SE even though they were marked SE. Therefore the
question to the list.
Thank you everyone for better educating me on the drives and how they should
work. I really appreciate it! It just goes to show no matter how much you
read, there is always something more you can find out by asking!
Cat
An added note to this - Friday mid-night when I began to install these
enclosures and drives, one of the single ended SCSII controllers went bad
(another item from the table - I think Murphy is using this check list
also). As I am sure you all know. a bad SCSII controller can hang up your
whole system to the point that booting hangs up regardless if you don't need
that volume set to operate or those drives for any production - aagh! At
that point a wonderful gentleman (Sabo) from HP came out and replaced the
card. When my system was operational again - somewhere around 5:30 am -
I decided the rest of these changes were better left for a few weeks until
after our primary software upgrade is completed, our nursing staff have
gotten over the downtime and their tempers and I have had a vacation!
============================================
Catherine Litten,
Senior Programmer/Analyst ([log in to unmask])
Valley Presbyterian Hospital (www.valleypres.org)
15107 Vanowen Street, Van Nuys, Ca 91405-4597
Phone (818) 782-6600 x 2140 Fax (818) 902-5202
============================================
The thoughts, comments, and opinions expressed herein are mine and do not
reflect those of my employer, or anyone else.
============================================
-----Original Message-----
From: HP-3000 Systems Discussion [mailto:[log in to unmask]]On
Behalf Of Steve Dirickson
Sent: Saturday, June 30, 2001 2:38 AM
To: [log in to unmask]
Subject: Re: Narrow vs. Wide Drives
> I have a question on what is the physical difference between
> a SE and DF
> drive is it firmware, toggle switches or something else?
They use different, and incompatible, electrical signaling standards. SE
SCSI is basically TTL-level (+5V) signals referenced to ground. Differential
(which some now call "HVD", "High Voltage Differential", as distinct from
"LVD" which we'll get to in a bit) is similar to the old EIA current-loop
setups, in that both lines are driven, one positive and one negative,
typically on the order of 12V. Differential thus has a much greater noise
resistance, which is why it allows for longer cables or for more devices at
higher speeds on a cable compared to an SE setup. Both types support 10MHz
("Fast") and 20MHz ("Fast-20", also called "Ultra") data rates. The reason
that regular SE is called "Fast" is that the original SCSI spec was 1.5MHz
for async or 5MHz for synchronous, but only weenies like me who have been
doing SCSI for two decades have ever even seen a non-"Fast" device.
Differential devices are required to disable themselves if they find that
they are on a bus with SE devices, so mixing types won't break anything (in
the absence of a hardware failure), but the differential devices won't be
visible on the bus.
The almost-latest iteration of differential is a hybrid that uses TTL-level
differential signals. It is called, appropriately enough, "Low Voltage
Differential"->"LVD", and uses a 40MHz clock. The lower voltage means that
the cables can't be as long as for HVD setups (12 meters for LVD as opposed
to 25m for HVD, except for point-to-point setups), but it allows for mixed
SE/LVD systems: if there are SE devices on the bus, the LVD devices operate
in SE mode (and at the lower 20MHz data rate). LVD is "almost-latest"
because the latest iteration is "Ultra160", which is just LVD at double the
speed: 80MHz x 16 bits = 160 MBytes/sec; it has the same SE compatibility as
LVD.
[Actually, the absolute "latest-latest" is Ultra320, but you can't buy an
Ultra320 setup today, and you'd need PCI-X to run it if you could]
> Whenever I have tried to install a drive that is marked SE part number
> A3646A is shows up in MAPPER as ST34572WC or ST34573WC
> (which are what I
> have in my production enclosures marked DF and attached via
> FASTWIDE cards).
For Seagate drives, a "W" on the right side of the size means "Wide", i.e. a
16-bit bus with a 68-pin connector (as opposed to "non-wide" drives,
indicated by "N", which use an 8-bit bus on a 50-pin connector). Unless
there's also a "C"; this indicates an SCA ("Single Connector Attachment")
drive, which uses an 80-pin connector that also carries the power and SCSI
ID lines. Functionally a "C" drive is identical to its non-"C" counterpart;
in fact, you can buy adapters to convert SCA drives for use on standard
68-conductor cabling. SCA drives are always "wide". "L" out there indicates
LVD, and "V" means Ultra160. "F" indicates a Fibre Channel drive. HP used to
support FC-AL on the 9x7 and 9x9 boxes; don't know about current machines.
So, your "WC" drives are SE drives with an SCA connector. Which sounds like
a problem, since HP's "FW" SCSI card is actually "FWD": "D" as in
"Differential", of the old variety. However, the A3646 part number is a
4.3GB Differential drive; maybe MAPPER is chopping off the "D" from the end?
(it's done it before)
> Although I can see them with MAPPER, Volutil says they are
> not useable no matter what I try.
>
> In reading further in my enclosures books it says both the DF
> and the SC are
> wide drives but I am assuming that it is a wide connection
> that the drive
> plugs into in the HASS, the difference in connecting to the
> Narrow Channel
> SCSII cards and WIDE to NARROW cable (thank you for the help
> understanding this last time).
I'm totally confused. What is connected to what? The HASS enclosure has two
independently-cabled halves, and each half can be SE or FWD (or you can
chain the two halves together), but the drives in each half must all be the
same type and must be connected to the proper controller (and terminated
with the proper terminator). If the drives are actually differential as
indicated by the part number, they should be connected to the FWD
controller. Don't get hung up on the wide/narrow issue; the make-or-break is
matching SE with SE and differential with differential.
Have you done the analysis in Table 10 in the HASS manual?
If MAPPER can see the drives, you should be OK electrically. In which case
I'd look at the SYSGEN config and see if it matches the devices found by
MAPPER and enumerated during system startup.
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
