1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
|
Frequently Asked Questions List, v 1.0.1
Index:
I. Compiling
II. Finding the correct device
III. Operational Issues
Part I: Compiling.
Q: Where is the Makefile in the tarball?
A: MTX now uses GNU Autoconf to generate the Makefile. Type "./configure"
while in the extracted mtx directory.
Q: Typing 'make' gives me a bunch of errors in the Makefile. Why can't
you provide a Makefile that works?
A: Note that you need the GNU 'make'. The BSD 'make' won't work, and
Solaris 'make' probably won't work either. If you want a better
configuration and makefile system, write one, then EMAIL me the results --
mtx is Open Source software and needs your code contributions to grow.
Q: How do I compile for operating systems other than Linux?
A: MTX no longer needs you to edit the Makefile to compile for operating
systems other than Linux. Just type ./configure and go with it.
Q: How do I port it to OS's other than the supported ones?
A: Create a new scsi_ module using one of the existing modules as an
example (scsi_freebsd.c might be a good model). Decide what symbol
you want #ifdef'ed in order to include that scsi_ module. Edit
mtxl.c to #include your scsi_ module. Edit the Makefile to add the
new target, including the -D needed to #include your new scsi_ module.
*********************************************************************
Part II: Finding the correct device.
Q: Why does this command not work??
[root@Scotty mtxl-1.4.8]# ./mtx -f /dev/st0 inquiry
In /var/log/messages I see:
st0: Write not multiple of tape block size.
A: Note that mtx 1.2 and above use the SCSI GENERIC interface on Linux,
FreeBSD, and Solaris (at least). They do NOT use the tape device node.
Q: When I do 'mtx -f /dev/sga inquiry' it shows
Product Type: Tape Drive
Vendor Id: HP
Product ID: C1553A
But when I do a 'mtx -f /dev/sga status' it fails. Why?!
A: You're trying to send a robotics command to a tape drive. You need
to send robotics commands to robotics devices, not to tape drives. Look in
lsscsi -gc (Linux) or camcontrol (FreeBSD) to find out what the
robotics device is. It will be reported as a 'Medium Changer', not a
'Sequential Access' or 'Tape Drive'.
Q: When I do 'cat /proc/scsi/scsi' it shows only one device, the tape device!
A: You are using a DAT autochanger that has one SCSI ID but two LUN's, LUN 0
and LUN 1. You need to compile a new kernel with SCAN SCSI LUNS enabled
or add this line to your /etc/lilo.conf (then run /sbin/lilo and reboot):
append="max_scsi_luns=2"
Q: I'm tired of typing '-f /dev/sg2' all the time. How do I set a default
device that 'mtx' looks at?
A: Set the CHANGER environment variable. For example, with 'bash':
export CHANGER=/dev/sg2
Q: I get "modprobe: can't locate module char-major-21"
syslog messages being squirreled away into a file on our syslog host,
and mtx doesn't work. What's the problem?
A: You need to compile SCSI generic support into your kernel (or as a module).
Q: When I installed mtx, a message showed
up on the console stating that a scsi changer was found at
dev sgr. However, I have no device /dev/sgr.
A: On Linux, device nodes should be created automatically by udev.
If you're not using udev, you could try 'mknod /dev/sgr c 21 19'
to create a device node. By default (if you're not using udev)
only 16 SCSI generic nodes are created. That might not be enough if
you have multiple SCSI controllers with lots of devices.
******************************************************
III. Operational issues:
Q: I'm using Red Hat 7.0 and mtx works fine when I type ./mtx from the
command line, but when I use it from scripts I get the following:
mtx: Request Sense: Error Code=70 (Current)
mtx: Request Sense: Sense Key=Aborted Command
mtx: Request Sense: Additional Sense Code = 4E
mtx: Request Sense: Additional Sense Qualifier = 00
What's happening?
A: Do "rpm -e mtx". Red Hat 7.0 includes a busted version of mtx. Your
script is apparently picking up the busted mtx in your path. Get rid
of the busted mtx, make sure that /usr/local/bin (or wherever you
put the "good" mtx) is in the path, and all should be well.
Q: I get
# /usr/local/bin/mtx -f /dev/sgr status
mtx: Request Sense: Error Code=70 (Current)
mtx: Request Sense: Sense Key=Not Ready
mtx: Request Sense: Additional Sense Code = 04
mtx: Request Sense: Additional Sense Qualifier = 8E
mtx: READ ELEMENT STATUS Command Failed
What gives?
A: Make sure your loader is in random mode, not sequential mode.
Most "real" loaders (as vs. DAT autoloaders) will not properly report
status information unless they are in "random" mode.
Q: I issue 'mtx load 5' and it loads tape 5. But when I try to put the tape
back in the magazine, we hit problems:
mtx: MOVE MEDIUM from Element Address 82 to 5 Failed
What gives?
A: Many loaders require you to first eject the tape (using 'mt' or 'tapectl')
before you issue an 'unload' command via 'mtx'.
Q: My Breece Hill loader does not properly report its slots.
A: Either set the "auto-inventory" feature in the loader's control panel,
or run 'mtx inventory' prior to running 'mtx status'.
Q: My Breece Hill loader takes a long time to do an inventory. mtx times
out and spits all over the place. Help!
A: Many loaders that support barcodes will perform poorly if you place tapes
into them without bar codes. Place bar codes on all your tapes and you
should be able to run 'mtx inventory' without that failure. You can
print your own bar codes (e.g. using http://tapelabel.de/).
Q: How do I eject the magazine of my autoloader?
A: Many low-end DAT autoloaders support the removable media 'EJECT' command
sent to the robotics device, even though it's not documented (or required)
in the SCSI standards. If the loader is at /dev/sg1, simply do
'mtx -f /dev/sg1 eject' and see what happens. (If nothing happens,
your autoloader doesn't support 'eject'). Some high-end libraries have
their own proprietary way for ejecting magazine trays, generally
involving abuse of the 'transfer' command and 'eepos' addendums,
but this is totally non-standard and undocumented.
Q: Is there a standard for cleaning tape bar codes?
A: Many libraries, and many backup programs, expect cleaning tape bar
codes to start with "CLN".
Q: How do I report a bug?
A: First, read this FAQ. Next, check the mtx list archives at
http://mtx.sourceforge.net to make sure that it's not already addressed
by somebody else. If your problem is still not solved, send
(to the mtx list) the following information:
Result of 'mtx inquiry' on the loader,
Result of 'mtx status' on the loader (minus a bunch of tapes if
it's a 50+ tape loader!),
Results of the operation that isn't working correctly.
|
