Re: SCSI Tape Drive Problems

2006-12-27 Thread Mark Kane
On Wed, Dec 20, 2006, at 11:14:57 -0600, Dan Nelson wrote:
 In the last episode (Dec 20), Mark Kane said:
  Thanks very much for the replies. Both drives are external, and they
  both have terminators on the back. The setup is like this:
  
  The DDS autoloader is connected to the back of the Adaptec card via
  a new external HD68 cable and the terminator that came with the
  drive is on the second plug of the unit.
  
  The Sun DLT drive is connected via an external HD50 cable to an
  adaptor that turns it into an internal cable which is then
  connected to the Adaptec card internally. It has a Sun terminator
  on the second plug on the unit.
  
  The cable that connects the DDS autoloader is brand new. It's not a
  major brand name, but it's listed as double shielded, UL20276
  listed, etc. The cable that connects the DLT unit is an older
  Adaptec one (possibly used) purchased for a few dollars locally. I
  could understand if that cable was a problem for writing using that
  drive, but having the issue on both drives like this with two
  separate cables just seems like it's something else.
  
  By the way, the green LEDs on both terminators are illuminated so
  they should be working.
 
 All that looks okay to me.  Try reposting your question to the
 [EMAIL PROTECTED] list.

Thanks again for the responses. Before I bug scsi@, I have a little
more information now that might help. I'll try to keep it simple.

When using pax, both drives write more data and get further than before
with cpio. The DLT (which previously gave those SCSI errors after
writing ~200 or ~500MB) can now write the full 17GB directory fine with
pax and without any SCSI errors or tape full problems. If I try to add
my 1.6GB mail directory onto that, it errors out with tape full after
about half my mail directory...so for simplicity's sake, it stops at
around 18GB with a tape full error.

On the DDS drive, pax gets considerably farther than cpio but still
cannot write the full 17GB directory. With pax, it errors out with a
tape full error pretty close to the end of that 17GB, but will never
complete it.

The files themselves inside those directories are of varying sizes,
however no single file is over 2GB.

So now I think it's pretty much down to drives saying the tape is full
before it really is without any specific SCSI errors anymore.

-Mark

-- 
Internet Radio:
Party107 (Trance/Electronic) - http://www.party107.com
Rock 101.9 The Edge (Rock) - http://www.rock1019.net

IRC:
MIXXnet IRC Network - irc.mixxnet.net (Nick: MIXX941)
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: SCSI Tape Drive Problems

2006-12-20 Thread Mark Kane
On Tue, Dec 19, 2006, at 23:16:45 -0600, Dan Nelson wrote:
 Timeouts and unexpected busfree errors like these are indicative of
 cabling or termination problems.  I don't think DLTs auto-terminate,
 for example, so try putting an external terminator on the 2nd plug on
 the back of the unit (or if it's an internal drive with one plug,
 terminate the cable the drive's plugged into).  Once you get those
 errors sorted out, see whether you still get incorrect volume full
 messages.

Thanks very much for the replies. Both drives are external, and they
both have terminators on the back. The setup is like this:

The DDS autoloader is connected to the back of the Adaptec card via a
new external HD68 cable and the terminator that came with the drive
is on the second plug of the unit.

The Sun DLT drive is connected via an external HD50 cable to an adaptor
that turns it into an internal cable which is then connected to the
Adaptec card internally. It has a Sun terminator on the second plug on
the unit.

The cable that connects the DDS autoloader is brand new. It's not a
major brand name, but it's listed as double shielded, UL20276
listed, etc. The cable that connects the DLT unit is an older Adaptec
one (possibly used) purchased for a few dollars locally. I could
understand if that cable was a problem for writing using that drive, but
having the issue on both drives like this with two separate cables just
seems like it's something else.

By the way, the green LEDs on both terminators are illuminated so they
should be working.

With the DDS drive, it does not give any errors to /var/log/messages
like the DLT drive does. cpio just quits with that Internal overflow,
aborting error (which doesn't happen with the DLT). Do you guys know
specifically what that message means? With the DLT drive, cpio actually
gives the volume full errors but with the DDS it's only Internal
Overflow and nothing else anywhere that I can see. Internet searches
for the Internal Overflow message have not turned up much helpful
information.

Thanks again.

-Mark

-- 
Internet Radio:
Party107 (Trance/Electronic) - http://www.party107.com
Rock 101.9 The Edge (Rock) - http://www.rock1019.net

IRC:
MIXXnet IRC Network - irc.mixxnet.net (Nick: MIXX941)
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: SCSI Tape Drive Problems

2006-12-20 Thread Derek Ragona
You can get ringing on a SCSI device from bad or loose cabling which will 
generate a lot of SCSI errors.  There are limits to cable lengths and you 
need to be sure you use proper cables for every device (and rated for the 
SCSI bus speed.)


You will do best to pull out any un-necessary hardware, use a minimal 
hardware setup.  You should also run diagnostics from the drive 
manufacturer and verify the tape and drive.  Last thing be sure to erase 
the tape before testing it.


-Derek


At 02:09 AM 12/20/2006, Mark Kane wrote:

On Tue, Dec 19, 2006, at 23:16:45 -0600, Dan Nelson wrote:
 Timeouts and unexpected busfree errors like these are indicative of
 cabling or termination problems.  I don't think DLTs auto-terminate,
 for example, so try putting an external terminator on the 2nd plug on
 the back of the unit (or if it's an internal drive with one plug,
 terminate the cable the drive's plugged into).  Once you get those
 errors sorted out, see whether you still get incorrect volume full
 messages.

Thanks very much for the replies. Both drives are external, and they
both have terminators on the back. The setup is like this:

The DDS autoloader is connected to the back of the Adaptec card via a
new external HD68 cable and the terminator that came with the drive
is on the second plug of the unit.

The Sun DLT drive is connected via an external HD50 cable to an adaptor
that turns it into an internal cable which is then connected to the
Adaptec card internally. It has a Sun terminator on the second plug on
the unit.

The cable that connects the DDS autoloader is brand new. It's not a
major brand name, but it's listed as double shielded, UL20276
listed, etc. The cable that connects the DLT unit is an older Adaptec
one (possibly used) purchased for a few dollars locally. I could
understand if that cable was a problem for writing using that drive, but
having the issue on both drives like this with two separate cables just
seems like it's something else.

By the way, the green LEDs on both terminators are illuminated so they
should be working.

With the DDS drive, it does not give any errors to /var/log/messages
like the DLT drive does. cpio just quits with that Internal overflow,
aborting error (which doesn't happen with the DLT). Do you guys know
specifically what that message means? With the DLT drive, cpio actually
gives the volume full errors but with the DDS it's only Internal
Overflow and nothing else anywhere that I can see. Internet searches
for the Internal Overflow message have not turned up much helpful
information.

Thanks again.

-Mark

--
Internet Radio:
Party107 (Trance/Electronic) - http://www.party107.com
Rock 101.9 The Edge (Rock) - http://www.rock1019.net

IRC:
MIXXnet IRC Network - irc.mixxnet.net (Nick: MIXX941)
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
MailScanner thanks transtec Computers for their support.


--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
MailScanner thanks transtec Computers for their support.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: SCSI Tape Drive Problems

2006-12-20 Thread Dan Nelson
In the last episode (Dec 20), Mark Kane said:
 On Tue, Dec 19, 2006, at 23:16:45 -0600, Dan Nelson wrote:
  Timeouts and unexpected busfree errors like these are indicative of
  cabling or termination problems.  I don't think DLTs auto-terminate,
  for example, so try putting an external terminator on the 2nd plug on
  the back of the unit (or if it's an internal drive with one plug,
  terminate the cable the drive's plugged into).  Once you get those
  errors sorted out, see whether you still get incorrect volume full
  messages.
 
 Thanks very much for the replies. Both drives are external, and they
 both have terminators on the back. The setup is like this:
 
 The DDS autoloader is connected to the back of the Adaptec card via a
 new external HD68 cable and the terminator that came with the drive
 is on the second plug of the unit.
 
 The Sun DLT drive is connected via an external HD50 cable to an adaptor
 that turns it into an internal cable which is then connected to the
 Adaptec card internally. It has a Sun terminator on the second plug on
 the unit.
 
 The cable that connects the DDS autoloader is brand new. It's not a
 major brand name, but it's listed as double shielded, UL20276
 listed, etc. The cable that connects the DLT unit is an older Adaptec
 one (possibly used) purchased for a few dollars locally. I could
 understand if that cable was a problem for writing using that drive, but
 having the issue on both drives like this with two separate cables just
 seems like it's something else.
 
 By the way, the green LEDs on both terminators are illuminated so they
 should be working.

All that looks okay to me.  Try reposting your question to the
[EMAIL PROTECTED] list.
 
 With the DDS drive, it does not give any errors to /var/log/messages
 like the DLT drive does. cpio just quits with that Internal
 overflow, aborting error (which doesn't happen with the DLT). Do you
 guys know specifically what that message means? With the DLT drive,
 cpio actually gives the volume full errors but with the DDS it's only
 Internal Overflow and nothing else anywhere that I can see.
 Internet searches for the Internal Overflow message have not turned
 up much helpful information.

If you look at the source to cpio, you can see in copyout.c, that
message is printed if a sprintf'ed header is larger than cpio expected
it to be.  My guess is one of your files is over 10gb and the file size
overflowed its 11-digit field.

-- 
Dan Nelson
[EMAIL PROTECTED]
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


SCSI Tape Drive Problems

2006-12-19 Thread Mark Kane
Hi everyone. I'm experiencing some problems with two tape drives
attached to this machine. Here is the situation:

I first tried backing up 17GB of data using cpio to a six tape DDS-4
autoloader with a fresh 20GB/40GB Fuji tape. It stopped writing after
about 9GB and gave the following error message on the command line,
but didn't output anything to /var/log/messages:

Internal overflow, aborting

If I try to use tar instead of cpio and backup the same 17GB directory
to the same 20GB native tape, it will give a tape full error way before
completing. 

With that six tape autoloader, I have tried with two different brand new
DDS-4 20/40 tapes, and the unit itself is brand new from the factory
and has only written a few GB so far (mainly smaller files, which
seemed to work fine).

I then tried backing up the same 17GB directory to a Sun DLT 4000 single
tape unit with a 20GB/40GB tape using cpio. It wrote about 250MB and
then quit with a Found End of Volume error. It outputted a lot of
errors to /var/log/messages which can be found here:

http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_1.log

I did another test on the same Sun DLT 4000 with a smaller 1.5GB
directory and a different tape. This time it wrote about 550MB and then
gave the same Found End of Volume message as before. This time there
were even more errors in /var/log/messages:

http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_2.log

I also tested with the smaller 1.5GB directory to the DDS-4 autoloader,
however that did complete without error.

At this point I was running 6.2-PRERELEASE #18: Wed Nov 15 11:12:09
from last month, so today I cvsupped to RELENG_6 and rebuilt. 

Trying the same 1.5GB directory using the Sun DLT 4000 after the update
resulted in it writing about 200MB of data and then stopping with the
Found end of volume message again, as well as errors
in /var/log/messages (however not as many this time):

http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_3.log

Any suggestions on what to check or any ideas on how to proceed would
be great. I don't think it's both tape drives failing or bad tapes
because pretty much the same thing happens with two different drives
(one of which is completely new), and four different tapes (two of those
completely new). I suppose it could be a problem with the SCSI
controller (Adaptec 2940U2W), but I'd like to rule out any software
issues before having to buy any additional hardware.

Thanks very much in advance.

-Mark
Dmesg: http://tntpowerhost.com/mixx941/freebsd/dmesg.121906

-- 
Internet Radio:
Party107 (Trance/Electronic) - http://www.party107.com
Rock 101.9 The Edge (Rock) - http://www.rock1019.net

IRC:
MIXXnet IRC Network - irc.mixxnet.net (Nick: MIXX941)
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: SCSI Tape Drive Problems

2006-12-19 Thread Dan Nelson
In the last episode (Dec 19), Mark Kane said:
 Hi everyone. I'm experiencing some problems with two tape drives
 attached to this machine. Here is the situation:
 
 I then tried backing up the same 17GB directory to a Sun DLT 4000
 single tape unit with a 20GB/40GB tape using cpio. It wrote about
 250MB and then quit with a Found End of Volume error. It outputted
 a lot of errors to /var/log/messages which can be found here:
 
 http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_1.log

Timeouts and unexpected busfree errors like these are indicative of
cabling or termination problems.  I don't think DLTs auto-terminate,
for example, so try putting an external terminator on the 2nd plug on
the back of the unit (or if it's an internal drive with one plug,
terminate the cable the drive's plugged into).  Once you get those
errors sorted out, see whether you still get incorrect volume full
messages.

Completely unrelated, but it's interesting that the string SCBs in
those logs got translated to SCDel!  I've never seen
Backspace-Delete conversion like that before :)

-- 
Dan Nelson
[EMAIL PROTECTED]
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]


Re: SCSI Tape Drive Problems

2006-12-19 Thread Ted Mittelstaedt
That's hardware, most likely scsi termination issues or cabling issues.

I have seen these errors lots of times when using a high speed scsi
controller
on a tape drive.  Sometimes you can play with the settings on the
controller or flip switches on the drive to get things to work, sometimes
you can use better cables and get things to work, usually I just
go get an older, slower scsi controller.  There's a computer recycler
here that has stacks of aha 2940's  (not the ultra ones) for a couple
bucks a card.

Ted

- Original Message - 
From: Mark Kane [EMAIL PROTECTED]
To: freebsd-questions@freebsd.org
Sent: Tuesday, December 19, 2006 8:16 PM
Subject: SCSI Tape Drive Problems


 Hi everyone. I'm experiencing some problems with two tape drives
 attached to this machine. Here is the situation:

 I first tried backing up 17GB of data using cpio to a six tape DDS-4
 autoloader with a fresh 20GB/40GB Fuji tape. It stopped writing after
 about 9GB and gave the following error message on the command line,
 but didn't output anything to /var/log/messages:

 Internal overflow, aborting

 If I try to use tar instead of cpio and backup the same 17GB directory
 to the same 20GB native tape, it will give a tape full error way before
 completing.

 With that six tape autoloader, I have tried with two different brand new
 DDS-4 20/40 tapes, and the unit itself is brand new from the factory
 and has only written a few GB so far (mainly smaller files, which
 seemed to work fine).

 I then tried backing up the same 17GB directory to a Sun DLT 4000 single
 tape unit with a 20GB/40GB tape using cpio. It wrote about 250MB and
 then quit with a Found End of Volume error. It outputted a lot of
 errors to /var/log/messages which can be found here:

 http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_1.log

 I did another test on the same Sun DLT 4000 with a smaller 1.5GB
 directory and a different tape. This time it wrote about 550MB and then
 gave the same Found End of Volume message as before. This time there
 were even more errors in /var/log/messages:

 http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_2.log

 I also tested with the smaller 1.5GB directory to the DDS-4 autoloader,
 however that did complete without error.

 At this point I was running 6.2-PRERELEASE #18: Wed Nov 15 11:12:09
 from last month, so today I cvsupped to RELENG_6 and rebuilt.

 Trying the same 1.5GB directory using the Sun DLT 4000 after the update
 resulted in it writing about 200MB of data and then stopping with the
 Found end of volume message again, as well as errors
 in /var/log/messages (however not as many this time):

 http://tntpowerhost.com/mixx941/freebsd/sun_tape_errors_3.log

 Any suggestions on what to check or any ideas on how to proceed would
 be great. I don't think it's both tape drives failing or bad tapes
 because pretty much the same thing happens with two different drives
 (one of which is completely new), and four different tapes (two of those
 completely new). I suppose it could be a problem with the SCSI
 controller (Adaptec 2940U2W), but I'd like to rule out any software
 issues before having to buy any additional hardware.

 Thanks very much in advance.

 -Mark
 Dmesg: http://tntpowerhost.com/mixx941/freebsd/dmesg.121906

 -- 
 Internet Radio:
 Party107 (Trance/Electronic) - http://www.party107.com
 Rock 101.9 The Edge (Rock) - http://www.rock1019.net

 IRC:
 MIXXnet IRC Network - irc.mixxnet.net (Nick: MIXX941)
 ___
 freebsd-questions@freebsd.org mailing list
 http://lists.freebsd.org/mailman/listinfo/freebsd-questions
 To unsubscribe, send any mail to
[EMAIL PROTECTED]


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]