Help - SCSI host number vs. Dell servers

  • Follow


Hi,

We created a Linux SCSI low-level driver.  By using /proc/scsi/scsi or
/proc/partitions or /dev/scsi, I can see our driver takes host2 on Dell 1750
server even thought host1 doesn't seem being used, and host1 on Dell 1850.
Why the same low-level driver take host2 on 1750 but host1 on 1850?

I don't think the low-level driver can control which host# to take, but the
Linux SCSI sub-system does that when scan the host/bus/target/lun, correct?

Thanks,


T.


0
Reply l_x2828 (152) 9/18/2004 12:10:50 AM

  "L" <l_x2828@yahoo.com>,
  In a message on 18 Sep 2004 00:10:50 GMT, wrote :

"> Hi,
"> 
"> We created a Linux SCSI low-level driver.  By using /proc/scsi/scsi or
"> /proc/partitions or /dev/scsi, I can see our driver takes host2 on Dell 1750
"> server even thought host1 doesn't seem being used, and host1 on Dell 1850.
"> Why the same low-level driver take host2 on 1750 but host1 on 1850?
"> 
"> I don't think the low-level driver can control which host# to take, but the
"> Linux SCSI sub-system does that when scan the host/bus/target/lun, correct?

What else is on the 1750 system?  What sort of kernel params are you using? 
Does the 1750 box have an IDE CD-R/RW (or DVD-R/RW/RAM)?  Is the kernel
getting an 'ide-scsi=...' parameter?  What could be happening is if
there is a IDE CD-R/RW type device on the 1750 box and the
'ide-scsi=...' is being used, then SCSI host #1 is the ide-scsi psuedo
SCSI device and then SCSI host #2 is the actually SCSI controller.

Also, if the SCSI controller is dual channel (which is not uncommon on a
server box), each channel is considered a separate 'host'.  It is then
possible that you might have your SCSI disks in the 1750 plugged in the
Channel B (second channel) and your SCSI disks in the 1850 plugged into
Channel A (first channel).  Thus host 1 on the 1750 will have no disks
and the disks will be on host 2.

"> 
"> Thanks,
"> 
"> 
"> T.
"> 
"> 
">                                                                                                          

                                     \/
Robert Heller                        ||InterNet:   heller@cs.umass.edu
http://vis-www.cs.umass.edu/~heller  ||            heller@deepsoft.com
http://www.deepsoft.com              /\FidoNet:    1:321/153






                                                         
0
Reply heller (2935) 9/18/2004 12:42:33 AM


"Robert Heller" <heller@deepsoft.com> wrote in message
news:94b3c$414b8479$cb248f0$8743@nf1.news-service-com...
>   "L" <l_x2828@yahoo.com>,
>   In a message on 18 Sep 2004 00:10:50 GMT, wrote :
>
> "> Hi,
> ">
> "> We created a Linux SCSI low-level driver.  By using /proc/scsi/scsi or
> "> /proc/partitions or /dev/scsi, I can see our driver takes host2 on Dell
1750
> "> server even thought host1 doesn't seem being used, and host1 on Dell
1850.
> "> Why the same low-level driver take host2 on 1750 but host1 on 1850?
> ">
> "> I don't think the low-level driver can control which host# to take, but
the
> "> Linux SCSI sub-system does that when scan the host/bus/target/lun,
correct?
>
> What else is on the 1750 system?  What sort of kernel params are you
using?
> Does the 1750 box have an IDE CD-R/RW (or DVD-R/RW/RAM)?  Is the kernel
> getting an 'ide-scsi=...' parameter?  What could be happening is if
> there is a IDE CD-R/RW type device on the 1750 box and the
> 'ide-scsi=...' is being used, then SCSI host #1 is the ide-scsi psuedo
> SCSI device and then SCSI host #2 is the actually SCSI controller.
>
> Also, if the SCSI controller is dual channel (which is not uncommon on a
> server box), each channel is considered a separate 'host'.  It is then
> possible that you might have your SCSI disks in the 1750 plugged in the
> Channel B (second channel) and your SCSI disks in the 1850 plugged into
> Channel A (first channel).  Thus host 1 on the 1750 will have no disks
> and the disks will be on host 2.

Thanks for the info.  Another persone also told me this and I have verified
that 1750 does have a dual on-board channels, while 1850 has only one.

Thanks,

T.
--

> ">
> "> Thanks,
> ">
> ">
> "> T.
> ">
> ">
> ">
>
>                                      \/
> Robert Heller                        ||InterNet:   heller@cs.umass.edu
> http://vis-www.cs.umass.edu/~heller  ||            heller@deepsoft.com
> http://www.deepsoft.com              /\FidoNet:    1:321/153
>
>
>
>
>
>
>


0
Reply l_x2828 (152) 9/20/2004 2:49:13 PM

2 Replies
43 Views

(page loaded in 0.04 seconds)

Similiar Articles:













7/23/2012 8:26:58 PM


Reply: