PDA

View Full Version : mtx -f /dev/sg3 hangs



alcina
October 14th, 2011, 08:37 AM
Hello,

Linux mt/mtx query - I'm not familiar enough with them to know what I'm looking at. I've just acquired a Quantum Scalar i80 with two LTO5 drives which is connected to the Ubuntu 10 server via SAS (a LSI Fusion-MPT HBA). At this precise moment I'm struggling to get mtx to give a response on the i80, yet the OS appears to be seeing it (there is also single LTO4 drive built in to the HP box):

# lsscsi -g
[0:0:0:0] disk ATA GB0160EAPRR HPG2 /dev/sda /dev/sg0
[3:0:0:0] tape HP Ultrium 5-SCSI Z3CZ /dev/st0 /dev/sg2
[3:0:0:1] mediumx QUANTUM Scalar i40-i80 130G /dev/sch0 /dev/sg3
[3:0:1:0] tape HP Ultrium 5-SCSI Z3CZ /dev/st1 /dev/sg4
[5:0:0:0] disk ATA Hitachi HDS5C302 ML6O /dev/sdb /dev/sg1
[7:0:0:0] tape HP Ultrium 4-SCSI U52D /dev/st2 /dev/sg5

All well and good. But... running "mtx -f /dev/sg3 inquiry" just hangs.

Running mtx inquiry on the two LTO5 drives produces what looks like sensible output (should the Attached Changer API line say "No" though?):
# mtx -f /dev/sg4 inquiry
Product Type: Tape Drive
Vendor ID: 'HP '
Product ID: 'Ultrium 5-SCSI '
Revision: 'Z3CZ'
Attached Changer API: No

# mtx -f /dev/sg2 inquiry
Product Type: Tape Drive
Vendor ID: 'HP '
Product ID: 'Ultrium 5-SCSI '
Revision: 'Z3CZ'
Attached Changer API: No

And running mt on sg3 produces:
# mt -f /dev/sg3 status
/dev/sg3: Device or resource busy

Nothing appears in /var/log/messages.

Any advice where to look?

Alcina

alcina
November 24th, 2011, 07:01 AM
Have found my own answer.....

This mysteriously "fixed itself" a couple of weeks later and I never got to the bottom of the problem. But it's happened again and as google kept finding this original post in my quest for a solution, I thought I'd update the thread so that the next time it happens I can google my own post again and lo the answer will be documented :D

The trick was to install sg_map (part of sg3-utils) which told me that /dev/sg6 (I now have more scsi devices and /dev/sg6 is now my changer) was busy.

# sudo sg_map
/dev/sg0 /dev/sda
/dev/sg1 /dev/sdb
/dev/sg2 /dev/sdc
/dev/sg3 /dev/sdd
/dev/sg4 /dev/sde
/dev/sg5 /dev/nst0
/dev/sg6 busy
/dev/sg7 /dev/nst1
/dev/sg8 /dev/nst2

So...what's keeping the changer busy?

# lsof /dev/sg6
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
dplinsdr 2119 root 29u CHR 21,6 0t0 2929 /dev/sg6

Aha! dplinsdr is part of the HP Data Protector software which was installed on this machine before I arrived and which I don't wish to use any more. Killing the dplinsdr process freed up /dev/sg6 and now "mtx -f /dev/sg6 status" gives a suitable response!

Huzzah! See you next google Alcina!