RE: Qualstar TLS-4220 problems
Oops sorry left the debug switches on from confirming the timeout loop. This
one will work better...
markh
-Original Message-
From: Mark Holm
Sent: Thursday, August 23, 2001 10:22 PM
To: 'Mitch Collinsworth'
Cc: '[EMAIL PROTECTED]'
Subject: RE: Qualstar TLS-4220 problems
Mitch,
Your hints were right on target. I converted your Perl to Bourne Shell and
finished converting the script to work with the Qualstar. It is now working
very well. Thank you! My box only has one drive, can anybody give me any
hints on how to make this work with two? Would I just define a second backup
set with a different set of drive parameters? Can somebody send me a mtx
status result from a Qualstar that has 2 tape drives?
I have attached the script as it stands for anybody else who has a Qualstar
with one drive.
markh
-Original Message-
From: Mitch Collinsworth [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, August 22, 2001 8:59 PM
To: Mark Holm
Cc: '[EMAIL PROTECTED]'
Subject: Re: Qualstar TLS-4220 problems
On Wed, 22 Aug 2001, Mark Holm wrote:
Qualstar TLS-4220 Tape Jukebox with 20 slots and 1 AIT-1 tape drive
It appears as if the changer mechanism and the tape drive are not
releasing the SCSI bus in a timely manner to allow the commands to
complete
when you switch devices.
Has anybody else seen this behavior? How do I fix it? Does anybody else
have
a Qualstar working with Amanda on RedHat 7.1?
I have a Qualstar TLS-4212 that I have not yet started interfacing
to Amanda, but will be doing so soon. What I have found with this
library, as well as with an Overland library I used in a previous
job, is that there is a delay between when a tape is loaded by the
library into the tape drive and when the tape drive becomes ready to
accept tape commands. Commands issued to the tape drive during this
period receive errors.
The Overland I used with a system that didn't at that time have a
supported changer script, so I wrote my own. I found I had to insert
a delay after loading a tape until the drive became ready before
returning to the calling program. Amanda tape changing worked
flawlessly in this manner.
I haven't yet begun setting up my Qualstar with Amanda but from
initial testing with a perl script it appears the same trick will be
necessary. Below is a copy of my test script, which illustrates how
I was able to get the Qualstar to behave itself under program control.
Note in particular the wait_for_drive_ready subroutine. This is stolen
more or less directly from my changer script for the Overland, except
for the open() test that worked the Overland's DLT drive but failed
to work with the Qualstar's AIT2 drive. This is replaced with a
rather kludgy (but working) routine called tape_online. I'd like to
think there's a better way to do this bit but in the short time I've
spent on it so far it hasn't come to me.
-Mitch
-
#!/usr/local/bin/perl
# qtest: some preliminary testing of qualstar tape library
# configuration data
$CHANGER = '/dev/sg4'; # device file for changer mechanism
$TAPE = '/dev/nst0';# device file for tape drive
$MT = /bin/mt -t $TAPE; # path to mt
$MTX = /sbin/mtx -f $CHANGER; # path to mtx
$MAXDELAY = 200;# max delay to insert after tape load
# to allow tape drive to become ready.
$FIRST_SLOT = 1;
$LAST_SLOT = 10;
#$slot = 10;
#while (1) {
# load_tape($slot++);
# if ($slot $LAST_SLOT) { $slot = $FIRST_SLOT; }
#}
while (1) {
$loaded_tape = tape_loaded;
print Current loaded tape: $loaded_tape\n;
next_tape;
}
exit 0;
#
# load_tape: load tape from specified slot
#
sub load_tape {
local($new_slot) = @_;
local($response,$result);
# first see if a tape is already loaded that needs to be put away
$loaded_slot = tape_loaded;
# put away loaded tape
if ($loaded_slot 0) {
if (tape_online) {
print ejecting tape from drive...\n;
system $MT offl;
}
print returning tape to slot $loaded_slot...\n;
system $MTX unload /dev/null 21;
}
# load requested tape
print loading tape from slot $new_slot...\n;
system $MTX load $new_slot;
# wait for tape loading to complete and drive to become ready
wait_for_drive_ready;
}
#
# next_tape: load next tape in sequence
#
sub next_tape {
local $loaded_slot = tape_loaded;
# if tape is in drive, eject first
if (tape_online) {
print ejecting tape from drive...\n;
system $MT offl;
}
# if current tape is from last slot, load tape from first slot
if ($loaded_slot == $LAST_SLOT) {
load_tape( $FIRST_SLOT );
} else {
# load next tape
print loading next tape\n;
system $MTX
RE: Qualstar TLS-4220 problems
On Thu, 23 Aug 2001, Mark Holm wrote: Your hints were right on target. I converted your Perl to Bourne Shell and finished converting the script to work with the Qualstar. It is now working very well. Thank you! Well, thank you! I was hoping to find time to tear into this next week and come up with a working changer script. Hopefully now I won't have to. I can just use yours! My box only has one drive, can anybody give me any hints on how to make this work with two? Would I just define a second backup set with a different set of drive parameters? I believe that's the general idea, yes. I'm wondering, though, what happens if one invocation of the changer script tries to issue a command to the changer device while the other invocation already has it busy. If the changer doesn't queue the requests but gives an error to the 2nd requestor, then presumably we'll need a separate process to drive the changer, and a changer script modified to send its requests to an input queue for that process. I _think_ there are at least a few folks here with multi-drive libraries. Anybody have any hard data to share? Can somebody send me a mtx status result from a Qualstar that has 2 tape drives? I can't today, but I have a 2nd drive on order so I should be able to answer this in a few days. I _believe_ we'll just see an additional Data Transfer Element line in mtx's output, same as adding an additional Storage Element. I expect this DTE will be numbered '1', and will not conflict numerically with Storage Element 1. But we'll see. -Mitch
Re: support for backup images larger than the tape drive
Hello, You wrote Can anybody tell me if amanda supports backup images larger than the tape drive. The readme file explicitly states that it does not, but then the file amanda-2.4.2p2/docs/Multitape talks about a need for this feature in version 2.2. No, it does not. The feature is being actively worked on, but no working code exists as of yet. I guess this has been on the Most Wanted list for quite some time :) Of course, the fact that it has been so long and it is still unimplemented gives a clue as to how difficult a job it is. If you are planning backup strategies for the near term, I would not count on multitape backup images for Amanda. I have complete confidence that it will be implemented, but I do expect that it will take a while for the Amanda developers to assure themselves that it works well enough to be called stable. -Ben -- Benjamin LewisThank goodness modern convenience is a Database Analyst/Programmer thing of the remote future. Purdue University Computing Center -- Pogo, by Walt Kelly [EMAIL PROTECTED]
RE: Qualstar TLS-4220 problems
Your welcome. I am now starting down the path of making it actually backup something and immediately stepped into the morass of tight permissions (even trying local), sigh! That's the hurdle for today... As to the conflict, it might be as simple as setting up another check for changer_online similar to the tape_online piece, or setup for checking against a lock file. Have to experiment. If it gets much more sophisticated, we will have to make this bourne script a Perl one, and I am not that good at Perl. Unless I miss my guess, when the second drive is added it will show up as 1 and the rest of the slots will move up one. MTX currently identifies the tape drive as slot 0, the holding slots as 1-20, the cleaning cartridge slots as 21-22, and the I/O port as 23. The only other option would be for the second drive to go to -1 and that doesn't make a whole lot of sense... markh -Original Message- From: Mitch Collinsworth [mailto:[EMAIL PROTECTED]] Sent: Friday, August 24, 2001 3:03 AM To: Mark Holm Cc: '[EMAIL PROTECTED]' Subject: RE: Qualstar TLS-4220 problems On Thu, 23 Aug 2001, Mark Holm wrote: Your hints were right on target. I converted your Perl to Bourne Shell and finished converting the script to work with the Qualstar. It is now working very well. Thank you! Well, thank you! I was hoping to find time to tear into this next week and come up with a working changer script. Hopefully now I won't have to. I can just use yours! My box only has one drive, can anybody give me any hints on how to make this work with two? Would I just define a second backup set with a different set of drive parameters? I believe that's the general idea, yes. I'm wondering, though, what happens if one invocation of the changer script tries to issue a command to the changer device while the other invocation already has it busy. If the changer doesn't queue the requests but gives an error to the 2nd requestor, then presumably we'll need a separate process to drive the changer, and a changer script modified to send its requests to an input queue for that process. I _think_ there are at least a few folks here with multi-drive libraries. Anybody have any hard data to share? Can somebody send me a mtx status result from a Qualstar that has 2 tape drives? I can't today, but I have a 2nd drive on order so I should be able to answer this in a few days. I _believe_ we'll just see an additional Data Transfer Element line in mtx's output, same as adding an additional Storage Element. I expect this DTE will be numbered '1', and will not conflict numerically with Storage Element 1. But we'll see. -Mitch
Re: support for backup images larger than the tape drive
In a message dated: Fri, 24 Aug 2001 10:06:25 CDT [EMAIL PROTECTED] said: Can anybody tell me if amanda supports backup images larger than the tape drive. No, it does not. The feature is being actively worked on, but no working code exists as of yet. Of course, the fact that it has been so long and it is still unimplemented gives a clue as to how difficult a job it is. By backup image, I'm interpreting that as a single file system. Is that correct? What about being able to have mutiple tapes per backup run? I thought I saw something about that mentioned recently. I thinks it's mentioned in the docs as well, that you can have a multi-tape run. Thanks, -- Seeya, Paul ...we don't need to be perfect to be the best around, and we never stop trying to be better. Tom Clancy, The Bear and The Dragon If you're not having fun, you're not doing it right!
unsubscribe amanda-users
unsubscribe amanda-users
Re: Qualstar TLS-4220 problems
Unless I miss my guess, when the second drive is added it will show up as 1 and the rest of the slots will move up one. MTX currently identifies the tape drive as slot 0, the holding slots as 1-20, the cleaning cartridge slots as 21-22, and the I/O port as 23. The only other option would be for the second drive to go to -1 and that doesn't make a whole lot of sense... In MTX, tape drives and cartridge slots are disjoint as far as element numbers are concerned. In other words, a second tape drive will show up as drive #1, but your slots will still be numbered from 1 to 20. Here's an example of a (mostly empty at the time) 2-drive library as reported by mtx: %mtx -f /dev/pass2 status Storage Changer /dev/pass2:2 Drives, 29 Slots ( 1 Import/Export ) Data Transfer Element 0:Full (Storage Element 4 Loaded):VolumeTag = DLT003 Data Transfer Element 1:Empty Storage Element 1:Full :VolumeTag=DLT000 Storage Element 2:Full :VolumeTag=DLT001 Storage Element 3:Full :VolumeTag=DLT002 Storage Element 4:Empty Storage Element 5:Full :VolumeTag=DLT004 Storage Element 6:Full :VolumeTag=DLT005 Storage Element 7:Full :VolumeTag=DLT006 Storage Element 8:Empty Storage Element 9:Empty Storage Element 10:Empty Storage Element 11:Empty Storage Element 12:Empty Storage Element 13:Empty Storage Element 14:Full :VolumeTag=DLT031 Storage Element 15:Empty Storage Element 16:Empty Storage Element 17:Empty Storage Element 18:Empty Storage Element 19:Empty Storage Element 20:Empty Storage Element 21:Empty Storage Element 22:Empty Storage Element 23:Empty Storage Element 24:Empty Storage Element 25:Empty Storage Element 26:Empty Storage Element 27:Empty Storage Element 28:Empty Storage Element 29 IMPORT/EXPORT:Empty Every other tape changer utility I've encountered starts numbering the storage elements at 0 too. Dunno why mtx starts at 1. -Ben -- Benjamin LewisThank goodness modern convenience is a Database Analyst/Programmer thing of the remote future. Purdue University Computing Center -- Pogo, by Walt Kelly [EMAIL PROTECTED]
amrecover problem
Hi all:
I've been backing up my system with amanda for the last week, and now
I'm doing the acid test: I'm trying to restore the backup to my laptop
computer, having upgraded to a larger disk. The backups seem to go
fine, according to the reports I've been getting. However, the restore
isn't going so well. I boot the laptop with tomsrtbt, having hacked
/etc/services to include the amanda, amandaidx, and amidxtape services.
Are the last two necessary?
When I start amrecover on the client, I get the following:
# amrecover -s {server} -t {server} -d {tape}
AMRECOVER Version 2.4.2ps. Contacting server on {server} ...
amrecover: amandaidx/tcp unknown protocol
Note that if I don't include all the switches above, it tries to contact
itself instead of the server. Have I screwed something up here?
One item to note is that I do *not* have inetd running; tomsrtbt doesn't
include it by default. There is a small amount of space left; on my
system, /sbin/inetd is 21K; if it doesn't need anything else, perhaps I
can include it. Will that solve my problem?
Thanks in advance,
John
RE: Qualstar TLS-4220 problems
Ben, Thanks. I see the difference now in the listing and should have before as well. The tapes are listed as Data Transfer Elements (DTE) and the slots are Storage Elements (SE), so the numbering does overlap some. The part I was overlooking is that the MTX program defaults to drive 0 (DTE 0) and that you can specify a DTE number as a second optional parameter to provide multi-drive targeting. I am going to have to dig into the source to see how it passes parameters to the changer script when two drives are defined. Experimentation is now in order, unless of course you are already using your Qualstar with Amanda and have a working script? markh -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Friday, August 24, 2001 9:44 AM To: Mark Holm Cc: 'Mitch Collinsworth'; '[EMAIL PROTECTED]' Subject: Re: Qualstar TLS-4220 problems Unless I miss my guess, when the second drive is added it will show up as 1 and the rest of the slots will move up one. MTX currently identifies the tape drive as slot 0, the holding slots as 1-20, the cleaning cartridge slots as 21-22, and the I/O port as 23. The only other option would be for the second drive to go to -1 and that doesn't make a whole lot of sense... In MTX, tape drives and cartridge slots are disjoint as far as element numbers are concerned. In other words, a second tape drive will show up as drive #1, but your slots will still be numbered from 1 to 20. Here's an example of a (mostly empty at the time) 2-drive library as reported by mtx: %mtx -f /dev/pass2 status Storage Changer /dev/pass2:2 Drives, 29 Slots ( 1 Import/Export ) Data Transfer Element 0:Full (Storage Element 4 Loaded):VolumeTag = DLT003 Data Transfer Element 1:Empty Storage Element 1:Full :VolumeTag=DLT000 Storage Element 2:Full :VolumeTag=DLT001 Storage Element 3:Full :VolumeTag=DLT002 Storage Element 4:Empty Storage Element 5:Full :VolumeTag=DLT004 Storage Element 6:Full :VolumeTag=DLT005 Storage Element 7:Full :VolumeTag=DLT006 Storage Element 8:Empty Storage Element 9:Empty Storage Element 10:Empty Storage Element 11:Empty Storage Element 12:Empty Storage Element 13:Empty Storage Element 14:Full :VolumeTag=DLT031 Storage Element 15:Empty Storage Element 16:Empty Storage Element 17:Empty Storage Element 18:Empty Storage Element 19:Empty Storage Element 20:Empty Storage Element 21:Empty Storage Element 22:Empty Storage Element 23:Empty Storage Element 24:Empty Storage Element 25:Empty Storage Element 26:Empty Storage Element 27:Empty Storage Element 28:Empty Storage Element 29 IMPORT/EXPORT:Empty Every other tape changer utility I've encountered starts numbering the storage elements at 0 too. Dunno why mtx starts at 1. -Ben -- Benjamin LewisThank goodness modern convenience is a Database Analyst/Programmer thing of the remote future. Purdue University Computing Center -- Pogo, by Walt Kelly [EMAIL PROTECTED]
Re: amrecover problem
John W. Price [EMAIL PROTECTED] wrote: amrecover: amandaidx/tcp unknown protocol something is missing in your /etc/services.
Re: Qualstar TLS-4220 problems
Mark Holm wrote: [...] I am going to have to dig into the source to see how it passes parameters to the changer script when two drives are defined. Experimentation is now in order, unless of course you are already using your Qualstar with Amanda and have a working script? Actually, my library is a BreeceHill, not a Qualstar. MTX shouldn't number the elements any differently for your library though. As far as changer scripts go, I've had good luck with the chg-scsi program and I've also used my own perl script. My OS (FreeBSD) provides its own changer tools though, so I've never bothered to try scripting mtx. Anyway, amanda doesn't pass any information about the tape drives to the changer script -- it actually wants the changer to provide the info back to it! You can use amgetconf from your script to query various parameters from the amanda.conf file if you want -- perhaps amgetconf tapedev. Since Amanda doesn't really care what the tapedev in amanda.conf is set to, I just set it to the drive number that the particular configuration should use and then do something like drivenum=`amgetconf tapedev` ... mtx -f $changerdev load $slotno $drivenum I use a config file to keep track of things like which device name is equivalent to drive 0. In a seperate message, you wrote: Storage Element 1:Full :VolumeTag=DLT000 Storage Element 2:Full :VolumeTag=DLT001 Storage Element 3:Full :VolumeTag=DLT002 Now this is interesting. Do you know anything about these VolumeTag fields? I haven't seen this before. Where are they stored such that mtx can get at them? Is amanda able to actually make use of them or does she still cycle through the library to read the tape labels? I.e. is 'amtape conf show' real fast or real slow? Those VolumeTags are the actual text from the barcodes on my tapes (printed my own). That info is stored by the library, from its barcode reader. Many changer scripts maintain some kind of database that maps the barcodes to the Amanda labels. Since I printed the barcodes myself, I cheated and made the barcodes and Amanda labels match. When Amanda (2.4.2+) uses a changer script/program, it first calls it with the -info option. One of the pieces of information Amanda is looking for is whether the changer is searchable. If the changer indicates that it is searchable, Amanda will ask it to load tapes by Amanda label rather than slot number if appropriate (if it knows it's looking for a particular label). You don't have to have a barcode reader to take advantage of this. If your changer script can keep some sort of database as to which tape is in which slot, you can implement the -search option. Much more information about this is in the file docs/TAPE.CHANGERS in the amanda source. You'll want to get the file from CVS since the searchable stuff wasn't documented when 2.4.2p2 was released (it was implemented though). You can also look at changer-src/chg-zd-mtx.sh.in -- that script takes advantage of those features. -Ben -- Benjamin LewisThank goodness modern convenience is a Database Analyst/Programmer thing of the remote future. Purdue University Computing Center -- Pogo, by Walt Kelly [EMAIL PROTECTED]
retry compiling on hp-ux 11i
Hi all. Posted this a few days ago but got no replies so I'll try again with the hope that my message just didn't get through the first time. I've compiled and been running amanda 2.4.2p2 on hp-ux 10.20 and 11.0 for some time with no probs. But now that we have a new file server with hp-ux 11i (11.11) I can't seem to get it to compile using gcc 3.0. Machine is an H-P A500 if that matters and I'm just trying to get the client side built. 'configure' runs ok with a few warnings about missing gnuplot, no readline lib, run amhpfixdevs, and, oddly, No working file locking capability found! Be VERY VERY careful. lockd and statd are running on system so I don't understand that last one. Anyway, the make completely fails with a warning about redefining __STDC_EXT__ and a series of parse errors pointing to various lines in sys/socket.h. I could post the exact output if needed. Thanks for any help. -- Larry Shanabrook --- [EMAIL PROTECTED] /\ || Bureau of Indian Affairs \/ Division of Energy and Mineral Resources __ 12136 W. Bayaud Ave /oo\ Suite 300 (|..|) Lakewood, CO 80228 \__/ (303)969-5270, ext 229 _--_
Re: Qualstar TLS-4220 problems
On Fri, 24 Aug 2001 [EMAIL PROTECTED] wrote: Storage Element 1:Full :VolumeTag=DLT000 Storage Element 2:Full :VolumeTag=DLT001 Storage Element 3:Full :VolumeTag=DLT002 Now this is interesting. Do you know anything about these VolumeTag fields? I haven't seen this before. Where are they stored such that mtx can get at them? Is amanda able to actually make use of them or does she still cycle through the library to read the tape labels? I.e. is 'amtape conf show' real fast or real slow? -Mitch
Re: Qualstar TLS-4220 problems
On Fri, 24 Aug 2001 [EMAIL PROTECTED] wrote: When Amanda (2.4.2+) uses a changer script/program, it first calls it with the -info option. One of the pieces of information Amanda is looking for is whether the changer is searchable. If the changer indicates that it is searchable, Amanda will ask it to load tapes by Amanda label rather than slot number if appropriate (if it knows it's looking for a particular label). Oh, cool! Now we can (better) support using multiple trays of tapes with a tape library, where we want to just stick in a new tray once per week or whenever. I like it! -Mitch
