Amanda-Users

RE: Minor problems with Spectralogic TreeFrog and chg-scsi

2003-04-16 18:26:28
Subject: RE: Minor problems with Spectralogic TreeFrog and chg-scsi
From: "Ron Bauman" <RBauman AT HatterasNetworks DOT com>
To: "Derek Suzuki" <DSuzuki AT ziprealty DOT com>, <amanda-users AT amanda DOT org>
Date: Wed, 16 Apr 2003 16:42:24 -0400
Derek, we're using a Treefrog 2000, Amanda-2.4.4 and Redhat Linux 7.3. Have 
been using the chg-zd-mtx script.  No problems such as those you've described.  
I did encounter the segfault you described just today.  We didn't have barcodes 
until now so I had no need to run update.  If you comment out the the statement 
"free(rest)" around line 537 in server-src/changer.c and recompile, you can get 
"amtape <conf> update" to work.

Here's my changer.conf file for reference:

# From /usr/local/src/amanda-2.4.4/changer-src/chg-zd-mtx
#
# Here is a commented out example file with all the variables and showing
# their default value (if any):
####
firstslot=1                 #### First storage slot (element) -- required
lastslot=12                 #### Last storage slot (element) -- required
cleanslot=15                #### Slot with cleaner tape -- default is "-1"
#                           #### Set negative to indicate no cleaner available
#
#   # Do you want to clean the drive after a certain number of accesses?
#   # NOTE - This is unreliable, since 'accesses' aren't 'uses', and we
#   #        have no reliable way to count this.  A single amcheck could
#   #        generate as many accesses as slots you have, plus 1.
#   # ALSO NOTE - many modern tape loaders handle this automatically.
#
AUTOCLEAN=1                 #### Set to '1' or greater to enable
autoclean=1                 #### Set to '1' or greater to enable
#
autocleancount=50           #### Number of access before a clean.
#
havereader=1                #### If you have a barcode reader, set to 1.
#
OFFLINE_BEFORE_UNLOAD=0 #### Does your tape driver require a
                        #### 'mt offline' before mtx unload?
offline_before_unload=0     #### Does your robot require an
#                           #### 'mt offline' before mtx unload?
#
driveslot=0                 #### Drive slot number.  Defaults to the
#                           #### tape device number (e.g. /dev/nst0 is
#                           #### assumed to be robot drive 0).
#
poll_drive_ready=1          #### Time (seconds) between tests to see if
#                           #### the tape drive has gone ready (default: 3).
#
max_drive_wait=120          #### Maximum time (seconds) to wait for the
#                           #### tape drive to become ready (default: 120).
unloadpause=0                   
loadpause=5                 #### Time (seconds) to wait after loading a tape
                            #### before declaring the operation done. (default: 
0)
offlinestatus=1           #### Set to 0 if 'mt status' gives an
#                         #### "offline" when drive is offline.
#                         #### Set to 1 or greater if 'mt status'
#                         #### doesn't give and offline, rather an
#                         #### "ONLINE" when drive is online.
####

Hope this helps.

Ron


-----Original Message-----
From: Derek Suzuki [mailto:DSuzuki AT ziprealty DOT com]
Sent: Tuesday, April 15, 2003 6:46 PM
To: 'amanda-users AT amanda DOT org'
Subject: Minor problems with Spectralogic TreeFrog and chg-scsi


        I have configured Amanda 2.4.4 on a Linux PC with an Adaptec
differential SCSI card and a Spectralogic Treefrog (AIT-2, 15-tape) changer.
My test rotation uses chg-scsi, with the changer.conf file attached at the
end of this message.  amtape segfaulted while I was labeling my tapes, but
the labels seemed to get recorded properly.  "amtape <conf> update"
segfaults while reading the first tape.
        Amanda seems to run okay if I just let it run unattended.  However,
if I try to tell the changer to switch to an arbitrary slot, the robot gets
confused about which tape goes into which real slot.  For example, it might
claim to be checking the tape in slot 2, but when the tape is unloaded it
will actually end up in slot 4.  Is this sort of thing normal?  I had hoped
to split the changer slots between two configurations, but that doesn't
appear to work at all.
        Meanwhile, I'm not sure whether the barcode reader is actually being
used.  Amanda does generate an accurate labelfile (though I didn't find a
standard way to turn that into text format), but it still sometimes seems
like it needs to load and read the tapes in order to identify them.
        Anyway, my backup rotation does seem to basically work.  I'm just
concerned about the changer confusion and wanted to see if this was a common
problem.

Derek Suzuki
dsuzuki AT ziprealty DOT com


# Configuration for chg-scsi tool
number_configs  1
eject           0
sleep           90
cleanmax        100
changerdev      /dev/sgb

# Configuration for drive #0
config          0
drivenum        0
dev             /dev/nst0
havebarcode     1
startuse        0
enduse          14
statfile        /var/lib/amanda/ziptest/tape0-slot
usagecount      /var/lib/amanda/ziptest/totaltime
tapestatus      /var/lib/amanda/ziptest/tapestatus
labelfile       /var/lib/amanda/ziptest/labelfile



<Prev in Thread] Current Thread [Next in Thread>