Hi all,

I have a redhat 7.1 server I am using as a TSM backup server. One of my VTLs started having issues yesterday on this box, and when I began digging into the problem, I found that this VTL and its six associated drives suddenly had duplicate device files (dmesg output below). I'm not sure how this happened, but it it's rendered that VTL unusable at the moment. Any ideas why this would happen? I haven't tried rebooting the box to see if the issue goes away. I'd like to find the answer to the problem before I do that.

My original devices, which were working fine:
[ 8745.794592] st 4:0:11:0: Attached scsi generic sg29 type 1
[ 8745.797998] st 4:0:12:0: Attached scsi generic sg30 type 1
[ 8745.800249] st 4:0:13:0: Attached scsi generic sg31 type 1
[ 8745.804094] st 4:0:14:0: Attached scsi generic sg32 type 1
[ 8745.807990] st 4:0:15:0: Attached scsi generic sg33 type 1
[ 8745.810183] st 4:0:16:0: Attached scsi generic sg34 type 1
[ 8745.819803] ch 4:0:17:0: Attached scsi generic sg35 type 8

And then these showed up:
859695.542297] st 4:0:11:0: Attached scsi generic sg29 type 1
[859695.543421] st 4:0:12:0: Attached scsi generic sg31 type 1
[859695.545937] st 4:0:13:0: Attached scsi generic sg32 type 1
[859695.550403] st 4:0:14:0: Attached scsi generic sg33 type 1
[859695.552568] st 4:0:15:0: Attached scsi generic sg34 type 1
[859695.557790] st 4:0:16:0: Attached scsi generic sg35 type 1
[859695.567843] ch 4:0:17:0: Attached scsi generic sg45 type 8

This VTL was using sg35, and now there's a virtual drive also assigned to that scsi device.

Any thoughts or insights would be welcomed.

Sincere thanks,
C.J.

Automatically assigned device names are not guaranteed to remain constant through reboots. You can use udev if you want persistent names based on (for example) SCSI address.

http://www-01.ibm.com/support/docvie...id=swg21459487