Single runlevel usb-ohci
 

I want to use my Seagate Travan USB tape drive (STT3401A) in single runlevel mode. It works perfectly in Slack runlevel 4 and also single runlevel on mandrake.

In slack the modules loaded in single are -
st, ide-scsi, nls_iso8859-15 and nls_cp850

and at runlevel 4 are (excluded sound and network)

st, parport_pc, lp, parport, usb-storage, usb-ohci, usbcore, pcmcia_core, ide-scsi, nls_iso8859-15 and nls_cp850

In single at command line I did
modprobe usb-ohci

the tape drive moves the tape and correctly identifies the STT3401A. I then tried

mt -f /dev/st0 eject

The tape was ejected but i got the error message
st0: Error with sense data: Current st09:00: sns=70 J5
ASC=20 ASQ= 0
Raw sense date: 0x70 0x00 0x05 0x00 0x00 0x00 ............................................................

I don't get this at slack runlevel 4, mandrake runlevel S or mandrake 6.

I obviously need to do something extra - but what?

AK

I would have thought that you'd need the usb-storage? I definitely need it for any of the usb storage devices I use (although I don't have tape)

Not got anywhere on this.

I've moved the hotplug code from /etc/rc.d/rc.M to /etc/rc.d/rc.K and booted the machine to runlevel 1 and the box recognises the tape drive and moves the tape but when I eject the tape I get the sense error. The tape ejects successfully. Modules usb-ohci, usb-storage are all loaded.

Don't get the error when I boot to runlevels 3 or 4 so obviously I need to do something more in rc.K other than the hotplug but I havn't a clue as to what it is.

The only thing to investigate is a line in my modules.conf which is

alias usb-controller usb-ohci

does the modules.conf file get actioned irrespective of the runlevel? I'm assuming it does.



AK

I'm no assuming until I find out otherwise that the error is not fatal. After all the tape ejects when told to irrespective of the error.

Therefore, I actioned a full backup of my system. Then restored 5 files from the backup to /tmp. The files were restored successfully.

Still would like to find out why it behaves differently at runlevels 1 and 4.

AK

Well I hadn't dug deeply enough on this one.

Although the tape device returns no error on Mandrake it does return the same error at irrespective of run levels in slackware. At runlevel 4 the error was being directed to /var/log/syslog.

In mandrake there looks like there are two extra modules that are loaded that may be significant - scsimon and scsi-mod. I don't know whether this matters or not as after all the tape drive is usable.

I suppose I should compare the differences in the way that the kernel is compiled between Mandrake and Slackware.

Does anyone know the purpose of scsimon and scsi-mod?

AK