| Author |
Message |
Jonathan Sachs
Guest
|
 Posted:
Sun Oct 10, 2004 3:51 am Post subject:
Puzzling problem with HP DAT drive & Adaptec host adapters |
|
|
I'm having a baffling problem with my backup device, a Hewlett-Packard DDS3
DAT drive. I hope someone can give me insight into what is happening, and
possibly suggest a cleaner solution than the one I have found.
The drive is mounted in an external enclosure. After purchasing it I used it
without problems on a Windows 2000 system for about 18 months, interfacing
it through the external connector on an Adaptec 29160N. I routinely verified
each backup after creating it, and got overall throughput of about
50MB/minute in each step.
In May I replaced the computer with a new one which runs Windows XP. Unlike
my previous computer this one does not have SCSI disk drives, so I got rid
of the 29160N and ran the DAT drive off an Adaptec 2910.
After a while I noticed a problem: the drive was still writing backups at
about 50MB/minute, but it was verifying at about 8MB/minute. (I originally
believed that this problem began a few weeks after I migrated to the new
machine. I no longer see how that could be so.)
I had a long dialogue with an expert in one of HP's user forums about the
problem. I ran a diagnostic utility and found that the verify step was
suffering massive read errors. He told me that my read/write heads were
either worn or misaligned. With his assistance, I ran a utility to realign
the heads, but it reported that the heads were already aligned correctly,
and the problem remained. We concluded that my drive was no longer working
correctly.
At this point a new person popped into the thread and told me that my
problem probably was not the drive at all, but the SCSI adapter. He said, in
essence: "The 2910 is a piece of junk which was actually designed for Apple
computers, not Wintel. It works well enough with older versions of Windows,
but Windows XP doesn't support it, as you would know if you had bothered to
check Microsoft's hardware compatibility list. Buy a real SCSI controller,
such as a 2930, and your problem will be solved."
I saw several inaccuracies in his claims, and did not give them much
credence. However, I remembered that my problems had begun about the time I
migrated to Windows XP, so I decided that his theory was worth testing. I
tried replacing the 2910 with a 2940AU, and got exactly the same results.
Then -- don't ask me why -- I tried running the drive off of an Adaptec
USB2Xchange. This is a gadget which runs a SCSI chain off a USB2 port. The
darned thing not only makes the drive work perfectly; it runs both the write
and verify steps at about 100MB/minute, twice as fast as before!
I am baffled twice. First, why is Windows XP having "issues" with the
2940AU, which is warranted compatible with it, as well as the 2910, which is
not? Second, why is the drive running twice as fast with the Adaptec
converter, which is basically a kludge that adds a second layer of hardware
and a second layer of drivers to the port, as it did with an adapter that
was designed to operate it?
I presented this puzzle in another forum, and was told that all of my
problems can be traced to either the cheap data cable or the cheap passive
terminator I must have been using with the SCSI cards. I found that
unpersuasive on several counts. First, I have used the data cable for years
with this drive and its predecessor, and never had a problem until I
migrated to Windows XP. Second, my terminator is from Granite Digital, which
is about as good as they get, and I used it for years without problems too;
in fact, I am using it with the USB2Xchange now.
I'd like to switch back to using a host adapter instead of the USB2Xchange
if I can figure out how to make it work without a 50% speed penalty on write
and a 92% penalty on verify. Any ideas what is happening? |
|
| Back to top |
|
 |
Folkert Rienstra
Guest
|
| Posted:
Mon Oct 11, 2004 10:52 pm Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
"Jonathan Sachs" <llm040903@earthlink.net> wrote in message news:5Q_9d.711$Bw1.202@newsread3.news.pas.earthlink.net
| Quote: | I'm having a baffling problem with my backup device, a Hewlett-Packard
DDS3 DAT drive. I hope someone can give me insight into what is happening,
and possibly suggest a cleaner solution than the one I have found.
The drive is mounted in an external enclosure. After purchasing it I used it
without problems on a Windows 2000 system for about 18 months, interfacing
it through the external connector on an Adaptec 29160N. I routinely verified
each backup after creating it, and got overall throughput of about 50MB/min.
in each step.
In May I replaced the computer with a new one which runs Windows XP.
Unlike my previous computer this one does not have SCSI disk drives, so
I got rid of the 29160N and ran the DAT drive off an Adaptec 2910.
After a while I noticed a problem: the drive was still writing backups
at about 50MB/minute, but it was verifying at about 8MB/minute.
(I originally believed that this problem began a few weeks after I
migrated to the new machine. I no longer see how that could be so.)
|
What changed that belief?
| Quote: |
I had a long dialogue with an expert in one of HP's user forums about the
problem. I ran a diagnostic utility and found that the verify step was
suffering massive read errors. He told me that my read/write heads were
either worn or misaligned. With his assistance, I ran a utility to realign
the heads, but it reported that the heads were already aligned correctly,
and the problem remained. We concluded that my drive was no longer working
correctly.
At this point a new person popped into the thread and told me that my pro-
blem probably was not the drive at all, but the SCSI adapter. He said, in
essence: "The 2910 is a piece of junk which was actually designed for Apple
computers, not Wintel. It works well enough with older versions of Windows,
but Windows XP doesn't support it, as you would know if you had bothered to
check Microsoft's hardware compatibility list. Buy a real SCSI controller,
such as a 2930, and your problem will be solved."
|
They are all based on the same chip design, the AIC-7850 and it's variations,
785x upto 7860.
| Quote: |
I saw several inaccuracies in his claims, and did not give them much credence.
However, I remembered that my problems had begun about the time I mi-
grated to Windows XP, so I decided that his theory was worth testing. I
tried replacing the 2910 with a 2940AU, and got exactly the same results.
|
Don't they both use the same driver?
| Quote: |
Then -- don't ask me why -- I tried running the drive off of an Adaptec
USB2Xchange. This is a gadget which runs a SCSI chain off a USB2 port.
The darned thing not only makes the drive work perfectly; it runs both the
write and verify steps at about 100MB/minute, twice as fast as before!
I am baffled twice. First, why is Windows XP having "issues" with the 2940AU,
which is warranted compatible with it, as well as the 2910, which is not?
Second, why is the drive running twice as fast with the Adaptec converter,
which is basically a kludge that adds a second layer of hardware and a
second layer of drivers to the port, as it did with an adapter that was
designed to operate it?
|
Definetely different driver.
| Quote: |
I presented this puzzle in another forum, and was told that all of my
problems can be traced to either the cheap data cable or the cheap passive
terminator I must have been using with the SCSI cards. I found that
unpersuasive on several counts. First, I have used the data cable for years
with this drive and its predecessor, and never had a problem until I migra-
ted to Windows XP. Second, my terminator is from Granite Digital, which
is about as good as they get, and I used it for years without problems too;
in fact, I am using it with the USB2Xchange now.
I'd like to switch back to using a host adapter instead of the USB2Xchange
if I can figure out how to make it work without a 50% speed penalty on write
and a 92% penalty on verify. Any ideas what is happening? |
|
|
| Back to top |
|
|
Jonathan Sachs
Guest
|
| Posted:
Tue Oct 12, 2004 5:16 am Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
"Folkert Rienstra" <see_reply-to@myweb.nl> wrote in message
news:2t063nF1q0mu4U1@uni-berlin.de...
| Quote: | (I originally believed that this problem began a few weeks after I
migrated to the new machine. I no longer see how that could be so.)
What changed that belief?
|
The facts that the problem turned out not to be a drive failure, and that
nothing else had changed since I migrated to the new system.
| Quote: | They are all based on the same chip design...
|
I'm not surprised. Just one more error in the information this guy was
giving me.
| Quote: | Don't they both use the same driver?
|
If they are all based on the same chip, very probably so.
| Quote: | Definetely different driver.
|
Do you think it's the driver? That would fit the observed facts, but if it's
so, I'm puzzled that no one has been able to point it out yet. I would
expect experts in the HP user forum to be aware if the Windows XP driver for
Adaptec's most populous family of adapters was incompatible with HP drives. |
|
| Back to top |
|
|
Ron Reaugh
Guest
|
| Posted:
Tue Oct 12, 2004 6:47 am Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
Cabling and/or termination is still the most probable source of the problem.
Remember that the SCSI chain must be terminated at both physical ends and
ONLY at the ends. Such a drive likely has onboard termination which may be
inadvertently enabled.
"Jonathan Sachs" <llm040903@earthlink.net> wrote in message
news:5Q_9d.711$Bw1.202@newsread3.news.pas.earthlink.net...
| Quote: | I'm having a baffling problem with my backup device, a Hewlett-Packard
DDS3
DAT drive. I hope someone can give me insight into what is happening, and
possibly suggest a cleaner solution than the one I have found.
The drive is mounted in an external enclosure. After purchasing it I used
it
without problems on a Windows 2000 system for about 18 months, interfacing
it through the external connector on an Adaptec 29160N. I routinely
verified
each backup after creating it, and got overall throughput of about
50MB/minute in each step.
In May I replaced the computer with a new one which runs Windows XP.
Unlike
my previous computer this one does not have SCSI disk drives, so I got rid
of the 29160N and ran the DAT drive off an Adaptec 2910.
After a while I noticed a problem: the drive was still writing backups at
about 50MB/minute, but it was verifying at about 8MB/minute. (I originally
believed that this problem began a few weeks after I migrated to the new
machine. I no longer see how that could be so.)
I had a long dialogue with an expert in one of HP's user forums about the
problem. I ran a diagnostic utility and found that the verify step was
suffering massive read errors. He told me that my read/write heads were
either worn or misaligned. With his assistance, I ran a utility to realign
the heads, but it reported that the heads were already aligned correctly,
and the problem remained. We concluded that my drive was no longer working
correctly.
At this point a new person popped into the thread and told me that my
problem probably was not the drive at all, but the SCSI adapter. He said,
in
essence: "The 2910 is a piece of junk which was actually designed for
Apple
computers, not Wintel. It works well enough with older versions of
Windows,
but Windows XP doesn't support it, as you would know if you had bothered
to
check Microsoft's hardware compatibility list. Buy a real SCSI controller,
such as a 2930, and your problem will be solved."
I saw several inaccuracies in his claims, and did not give them much
credence. However, I remembered that my problems had begun about the time
I
migrated to Windows XP, so I decided that his theory was worth testing. I
tried replacing the 2910 with a 2940AU, and got exactly the same results.
Then -- don't ask me why -- I tried running the drive off of an Adaptec
USB2Xchange. This is a gadget which runs a SCSI chain off a USB2 port. The
darned thing not only makes the drive work perfectly; it runs both the
write
and verify steps at about 100MB/minute, twice as fast as before!
I am baffled twice. First, why is Windows XP having "issues" with the
2940AU, which is warranted compatible with it, as well as the 2910, which
is
not? Second, why is the drive running twice as fast with the Adaptec
converter, which is basically a kludge that adds a second layer of
hardware
and a second layer of drivers to the port, as it did with an adapter that
was designed to operate it?
I presented this puzzle in another forum, and was told that all of my
problems can be traced to either the cheap data cable or the cheap passive
terminator I must have been using with the SCSI cards. I found that
unpersuasive on several counts. First, I have used the data cable for
years
with this drive and its predecessor, and never had a problem until I
migrated to Windows XP. Second, my terminator is from Granite Digital,
which
is about as good as they get, and I used it for years without problems
too;
in fact, I am using it with the USB2Xchange now.
I'd like to switch back to using a host adapter instead of the USB2Xchange
if I can figure out how to make it work without a 50% speed penalty on
write
and a 92% penalty on verify. Any ideas what is happening?
|
|
|
| Back to top |
|
|
Jonathan Sachs
Guest
|
| Posted:
Thu Oct 14, 2004 4:45 am Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
"Ron Reaugh" <rondashreaugh@att.net> wrote in message
news:5BHad.532182$OB3.331919@bgtnsc05-news.ops.worldnet.att.net...
| Quote: | Cabling and/or termination is still the most probable source of the
problem.
Remember that the SCSI chain must be terminated at both physical ends and
ONLY at the ends. Such a drive likely has onboard termination which may
be
inadvertently enabled.
|
Remember that the SAME TERMINATOR, DEVICE, and CABLE worked with no problems
under Windows 2000. It is not clear to me how cabling or termination could
be at fault when those factors have remained constant, while another factor
changed at the time when the problem appeared.
Do you remember what they taught about experimental design in your college
science courses? In its simplest form, a controlled experiment consists of
changing ONE factor in a situation, while keeping other factors constant. If
the change in the variable factor is accompanied by a change in the result,
that is ordinarily taken as evidence that the changed result is causally
related to the variable factor. I cannot think of a situation where it is
taken as evidence that the changed result is causally related to one of the
controlled factors.
Don't get me wrong. I'm not ruling out the possibility that the terminator
or cable is at fault. Equally strange things have been known to happen.
Superfluidity is one, although I can't think of any others at the moment.
But if that is the case, you're going to have to explain how it is possible. |
|
| Back to top |
|
|
Ron Reaugh
Guest
|
| Posted:
Thu Oct 14, 2004 5:38 am Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
"Jonathan Sachs" <llm040903@earthlink.net> wrote in message
news:a%jbd.3555$gy1.1052@newsread1.news.pas.earthlink.net...
| Quote: | "Ron Reaugh" <rondashreaugh@att.net> wrote in message
news:5BHad.532182$OB3.331919@bgtnsc05-news.ops.worldnet.att.net...
Cabling and/or termination is still the most probable source of the
problem.
Remember that the SCSI chain must be terminated at both physical ends
and
ONLY at the ends. Such a drive likely has onboard termination which may
be
inadvertently enabled.
Remember that the SAME TERMINATOR, DEVICE, and CABLE worked with no
problems
under Windows 2000.
|
Not relevant. Bad termination or cabling can rear its ugly head at its
choice of time and circumstance and choose to behave nicely at other times.
Maxwell made him do it.
| Quote: | It is not clear to me how cabling or termination could
be at fault when those factors have remained constant, while another
factor
changed at the time when the problem appeared.
|
Doesn't work that way.
| Quote: | Do you remember what they taught about experimental design in your college
science courses? In its simplest form, a controlled experiment consists of
changing ONE factor in a situation, while keeping other factors constant.
If
the change in the variable factor is accompanied by a change in the
result,
that is ordinarily taken as evidence that the changed result is causally
related to the variable factor.
|
You mean the logic that goes: There are more TVs in the US than Japan and
there is more cancer in the US than Japan so obviously cancer causes TVs.
| Quote: | I cannot think of a situation where it is
taken as evidence that the changed result is causally related to one of
the
controlled factors.
Don't get me wrong. I'm not ruling out the possibility that the terminator
or cable is at fault.
|
That's the high probability.
| Quote: | Equally strange things have been known to happen.
Superfluidity is one, although I can't think of any others at the moment.
|
How about that the neutrons in a neutron star can form a solid phase.
There's something about the "neutron drip point" below which it qualifies as
a solid.
| Quote: | But if that is the case, you're going to have to explain how it is
possible. |
W2K -> XP
29160N -> 2910
"I would expect experts in the HP user forum to be aware if the Windows XP
driver for Adaptec's most populous family of adapters was incompatible with
HP drives."
Have you heard of the process of elimination?
It's cabling and/or termination. Such a faulty states are known to work for
an extended period and then change one thing and all hell breaks loose. |
|
| Back to top |
|
|
Triffid
Guest
|
| Posted:
Thu Oct 14, 2004 9:05 am Post subject:
Re: Puzzling problem with HP DAT drive & Adaptec host adapte |
|
|
Jonathan Sachs wrote:
| Quote: | "Ron Reaugh" <rondashreaugh@att.net> wrote in message
news:5BHad.532182$OB3.331919@bgtnsc05-news.ops.worldnet.att.net...
Cabling and/or termination is still the most probable source of the
problem.
Remember that the SCSI chain must be terminated at both physical ends and
ONLY at the ends. Such a drive likely has onboard termination which may
be
inadvertently enabled.
Remember that the SAME TERMINATOR, DEVICE, and CABLE worked with no problems
under Windows 2000. It is not clear to me how cabling or termination could
be at fault when those factors have remained constant, while another factor
changed at the time when the problem appeared.
|
You have not considered the change process itself - which typically
involves powering down, unplugging and plugging, rearranging cables etc.
- all on equipment which may have run undisturbed for years. The
probability of component failure occuring during (or shortly after)
change is much higher than during normal operation.
It wouldn't be the first time a perfectly good SCSI cable gave trouble
after being moved from one system to another ;-)
| Quote: | Do you remember what they taught about experimental design in your college
science courses? In its simplest form, a controlled experiment consists of
changing ONE factor in a situation, while keeping other factors constant. If
the change in the variable factor is accompanied by a change in the result,
that is ordinarily taken as evidence that the changed result is causally
related to the variable factor. I cannot think of a situation where it is
taken as evidence that the changed result is causally related to one of the
controlled factors.
Don't get me wrong. I'm not ruling out the possibility that the terminator
or cable is at fault. Equally strange things have been known to happen.
Superfluidity is one, although I can't think of any others at the moment.
But if that is the case, you're going to have to explain how it is possible.
|
|
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in