| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 9564.1 | | SMURF::KNIGHT | Fred Knight | Tue Apr 22 1997 15:35 | 14 |
| Thats what happens when you buy a NON-SUPPORTED device
that violates the scsi spec.
The scsi-2 spec clearly states that the lun field is
a 3 bit field. So, I'm not sure how you can have any
lun number greater than 7 (atleast I can't think of
how to do it).
If other vendors drivers also violate the scsi-2 spec,
then so be it. We however comply with the standard and
will work correctly with devices that also comply with
the standard.
Fred Knight
|
| 9564.2 | | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Tue Apr 22 1997 18:47 | 4 |
| Ah, but would it violate the spec (such as it is) if the device
where claiming to be SCSI-3? SCSI-2 was as solid as quick-sand
when vendors started claiming "compliance". Why should they do
any differently for SCSI-3...
|
| 9564.3 | Boxhill & scsi compliance | SIOG::BR_MURPHY | | Wed Apr 23 1997 05:25 | 18 |
| Re 9564.1
Hey!! Don't blame me for buying this stuff. I was just responding to a
call & found it already installed.
Actually, I don't believe it's a none standard scsi2-spec. I think it
complies very well with the spec. I think the real problem is that the
VAR's who sold it & installed it don't really know how to configure it.
I was hoping there was someone out there who could help me sort out the
correct configuration...or point me in the right direction for further
information.
Re 9564.2
SCSI-3? Why what does the scsi3 spec have to say about LUNS, & does
Digital supply scsi cards that conform to this standard....& are
compatible with AS4100 running DU3.2G
|
| 9564.4 | | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Wed Apr 23 1997 18:36 | 5 |
| I'm lead to believe that SCSI-3 increases the number of logical
units that a target can have. I don't know if support for this
particular feature requires new adapters or just drivers and
supporting host software.� I'm also lead to believe that a
future version of Digital UNIX will be more SCSI-3 friendly.
|
| 9564.5 | | DECWET::RWALKER | Roger Walker - Media Changers | Tue Apr 29 1997 20:28 | 12 |
| SCSI-2, the current spec, limits the LUN to 3 bits in
the IDENTIFY message. This only allows a LUN range of
0-7. While SCSI-3 is being drafted some vendors are using
some of its featues and some even report SCSI-3 in the
inquire data. Why wait for a standard to finish if it
has good stuff, we couldn't for real time POSIX.
To use of the SCSI-3 LUN range requires all parts of the system
to change. Even if you could get an adapter that works
it doesn't fix the O/S. We almost had it for 4.0 but oh well.
Until all software can live without the /dev/rz5c format
we just have to do without more LUNS.
|
| 9564.6 | | SMURF::KNIGHT | Fred Knight | Wed Apr 30 1997 17:32 | 9 |
| The SCSI-3 lun field is 8 bytes in length. So yes, it
requires not only new S/W, but new H/W as well!
Also, any device that uses less than 8 bytes for the
lun field is NOT SCSI-3. And any device that uses more
than 3 bits is NOT SCSI-2. So, if it's not SCSI-2, and
it's not SCSI-3, .....
Fred Knight
|