How to configure EML103e using SLES9?

[Tom Kunkel]

We have an EML103e fibre attached (via SAN) to a SLES9 box.

In the /proc/scsi/qla2xxx/0 and 1 I can see the HBA's of the EML.

How can I get the library to be recognised by the OS?

[Alan_152]

Most tape devices should show up in /dev after rebooting...

[Ivajlo Yanakiev]

use dmesg to find all know device.

[Tom Kunkel]

Thanks but been there, done all that.

I can see the /dev/st0 and /dev/st1 drives. However I cannot seem to find a device for the EML library.

Any ideas?

[Uwe Zessin]

You need to configure the embedded E2400 "storage router" (=FC bridge) to map the EML's devices into the FC LUN address space of the Linux server.

[Uwe Zessin]

Ah, responses crossed ;-)

Do you see any "sg" devices?

[Tom Kunkel]

Yup:

# lssg
sg0 0,0,0,0 HP HSV200 5030 5000-1fe1-5007-6ff9
sg1 0,0,0,2 HP HSV200 5030 5000-1fe1-5007-6ff9
sg2 0,0,0,8 HP HSV200 5030 5000-1fe1-5007-6ff9
sg3 1,0,1,0 HP EML E-Series 1022 1000-00e0-0223-39e4
sg4 1,0,1,1 HP Ultrium 3-SCSI L29S 1000-00e0-0223-39e4
sg5 1,0,1,2 HP Ultrium 3-SCSI L29S 1000-00e0-0223-39e4


0 is a scsi raid device and 1 and 2 are FC luns.

What are you thinking?

[Uwe Zessin]

Looks like "sg3" is the robotic, "sg4" and "sg5" are the tape drives.

[Tom Kunkel]

Um yes, but usually (and here I'm thinking IBM3494 robots) there is a piece of software that drives the ATL and communicates with it at a much higher level.

I guess that I must be missing something I have yet to find out about.

Oh well. Back to the old drawing board.

[Tom Kunkel]

Found out what it was shortly after (but did not know how to find this old thread:

The EML library's firmware by default does not read barcodes properly. All barcode labels from HP are 8 characters long yet the robot's set up for 6.

Believe it or not, that was it.