[clug] SCSI information
Burn Alting
burn at goldweb.com.au
Thu Jan 6 11:50:31 GMT 2005
Michael,
What happens when you
a. cat /proc/scsi/scsi
then
b. add an additional lun (that you configured via the Apple gui)
using something like
echo "scsi add-single-device $bus 0 $sid $lun" > /proc/scsi/scsi
Where 'bus' is the bus number reported in a as is sid.
EG If I saw
[root at swtf3 host_utils]# cat /proc/scsi/scsi
Attached devices:
Host: scsi1 Channel: 00 Id: 00 Lun: 00
Vendor: SciComp Model: CX450-TP Rev: 0207
Type: Sequential-Access ANSI SCSI revision: 03
Host: scsi1 Channel: 00 Id: 00 Lun: 01
Vendor: SciComp Model: CX450-IMC Rev: 0207
Type: Medium Changer ANSI SCSI revision: 03
and I KNEW I had another lun at '2' then I would
echo "scsi add-single-device 1 0 0 2" > /proc/scsi/scsi
I get 1 from scsi1, the first 0 from Channel, the second from Id:
and the lun from the set lun on the device.
You may want to download Doug Gilbert's sg_utils (www.torque.net/sg)
his sg_map is very useful.
Regards
Burn Alting
burn at goldweb.com.au
On Thu, 2005-01-06 at 19:58, Michael Carden wrote:
> I'm looking for a good source of documentation on how a Linux kernel looks at
> SCSI devices.
>
> My situation is that I have some Apple Xserve RAID that I'm looking at via Red
> Hat ES3 (2.4 kernel). One issue already is the fact that RHES3 can't 'see'
> the default 2.18 terabyte volumes that Apple sets up as defaults, so no
> partitioning or filesystem creation can happen until the default volumes are
> destroyed and replaced with smaller ones. But I digress...
>
> Via Apple's fine Java-based[1] admin utility, I can create various LUNs from
> the RAID arrays. The LUNs exist and may be seen via the admin app or via the
> Fibre Channel HBA app that declares what's hanging off the HBA, but... I
> can't get the kernel to recognise anything more than LUN0 on either side of
> the Xserve RAID.
>
> The setup is:
>
> Dell PowerEdge 2500 with RHES3
> Emulex HBA with LC-connected fibre to;
> Brocade 3850 FC Switch
> Apple 5600G Xserve RAID with two SFP/FC transceivers and fibre to the switch.
>
> Now I accept that this comes about through my ignorance of the underlying
> processes here, so I have been studiously googling for clues but so far I
> have failed to uncover the secrets I need.
>
> Any suggestions?
>
> Thanks,
> MC
>
>
> [1] Looks fine under Java 1.4.2 but text boxes are broken under 1.5... er...
> Java 5...
> --
> http://michaelcarden.net
More information about the linux
mailing list