Dear,
I'm trying to connect some FC Tape drives to a Linux system and I'm having some difficulties.
first of all, I'm runing Linux version 2.6.9-89.ELsmp (Red Hat 3.4.6-11)

My FC HBA is well recognized:
kernel: SCSI subsystem initialized
kernel: Emulex LightPulse Fibre Channel SCSI driver 8.0.16.46
kernel: Copyright(c) 2003-2008 Emulex. All rights reserved.
kernel: ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 18 (level, low) -> IRQ 201
kernel: lpfc 0000:00:0e.0: 0:1303 Link Up Event x1 received Data: x1 x1 x0 x2 x0 x0
kernel: lpfc 0000:00:0e.0: 0:1305 Link Down Event x2 received Data: x2 x20 x1000 x0 x0
kernel: lpfc 0000:00:0e.0: 0:1303 Link Up Event x3 received Data: x3 x1 x0 x2 x0 x0
kernel: scsi0 : Emulex LP7000 1Gb PCI Fibre Channel Adapter on PCI bus 00 device 70 irq 201

and I can see my tape drives correctly:
>cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: IBM Model: 03590E1A Rev: E32E
Type: Sequential-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 00 Lun: 01
Vendor: IBM Model: 03590E1A Rev: E32E
Type: Sequential-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 00 Lun: 02
Vendor: IBM Model: 03590E1A Rev: E32E
Type: Sequential-Access ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 00 Lun: 03
Vendor: IBM Model: 03590E1A Rev: E32E
Type: Sequential-Access ANSI SCSI revision: 03

And in the var.log.messages, the tests of the tape drives look OK:
kernel: Vendor: IBM Model: 03590E1A Rev: E32E
kernel: Type: Sequential-Access ANSI SCSI revision: 03
kernel: Vendor: IBM Model: 03590E1A Rev: E32E
kernel: Type: Sequential-Access ANSI SCSI revision: 03
kernel: Vendor: IBM Model: 03590E1A Rev: E32E
kernel: Type: Sequential-Access ANSI SCSI revision: 03
kernel: Vendor: IBM Model: 03590E1A Rev: E32E
kernel: Type: Sequential-Access ANSI SCSI revision: 03
kernel: st: Version 20040403, fixed bufsize 32768, s/g segs 256
kernel: Attached scsi generic sg0 at scsi0, channel 0, id 0, lun 0, type 1
kernel: Attached scsi generic sg1 at scsi0, channel 0, id 0, lun 1, type 1
kernel: Attached scsi generic sg2 at scsi0, channel 0, id 0, lun 2, type 1
kernel: Attached scsi generic sg3 at scsi0, channel 0, id 0, lun 3, type 1
kernel: Attached scsi tape st0 at scsi0, channel 0, id 0, lun 0
kernel: st0: try direct i/o: yes (alignment 512 B), max page reachable by HBA 4294967295
kernel: Attached scsi tape st1 at scsi0, channel 0, id 0, lun 1
kernel: st1: try direct i/o: yes (alignment 512 B), max page reachable by HBA 4294967295
kernel: Attached scsi tape st2 at scsi0, channel 0, id 0, lun 2
kernel: st2: try direct i/o: yes (alignment 512 B), max page reachable by HBA 4294967295
kernel: Attached scsi tape st3 at scsi0, channel 0, id 0, lun 3
kernel: st3: try direct i/o: yes (alignment 512 B), max page reachable by HBA 4294967295

The issue that I have is when I try to access those devices, I've got the following error:
>mt -f /dev/st0 status
/dev/st0: Input/output error
Same error occured if tape is loaded or drive empty.

I can see in the /var/log/messages
kernel: scsi0 (0,0,0) : reservation conflict

Do you have any idea of the problem ?
How to solve it ?

Thanks very much for your help.

Best regards,

Reservation conflict means that the tape drives are in use by another device. Once that device is done it should release the reservation.

Thanks for your help.
But the drives are direct attached to the Linux server... so it is impossible that another device has reserve those tape drives !!!
After connecting the tape drives to the Linux, should I run a command or something else to make the drives available ?

Thanks one more time for your support.

Best regards,

This is RHEL 3.4 ? right
the very old 3.4
the no longer supported RHEL3

did you ask your red hat rep ?

No, no... we are using RedHat $ Update 8.

Regards,

Why do you say it is impossible? Just because the device is no longer connected does not mean that it did not set a reservation in the past. Perhaps you moved the cable without properly shutting down the application. You can clear a reservation using sg3_utils (either sg_reset or sg_persist depending on the reservation type).