External HDD Issue

  • Thread starter Thread starter Nefrit
  • Start date Start date
N

Nefrit

I've got a SQL server that uses an external HDD to do backup/transaction
logs. The drive specifically has to be the "I" drive in order for the
scripts to run.

All was working fine untill I replaced the 5 external drives with new
ones. Same size, same brand, different enclosure.

Now whenever I plug the drives in, they do not get an assigned drive
letter. The only way the server will see them is if I go into "Manage
your computer" and then assign the drive the proper letter.

This is one of those: I've never seen that before.

Do you guys have any ideas?
_________________
'Home Mortgage Refinancing'
(http://www.mortgagerefinancing-central.com) 'Make up Games'
(http://www.dressupwho.com/categories/5/makeovers.html)




--
Nefrit
 
If I: isn't the first free local drive letter then the
old five drive must have had completely identical USB
vendor and device IDs and identical USB serial numbers.

Therefore XP sees all five drives as the same one and
assigns the same drive letter.

Identical USB hardware serial are a violation of the
USB specs. Two of these drive will not work at the
same time offhand.

Gettings the same letter for non indentical drives
only works on the first free letter. The drive letter
assignments for the drives overwrite each other then
and XP assings the first free letter each time.

For a high letter you need a 3rd party software as
my USB drive letter manager 'USBDLM':
http://www.uwe-sieber.de/usbdlm_e.html

It's a Win32 service that gives control over drive
letter assingments, primarily of USB drives.

To get USB hard drives at I: and USB flash drives
at U: to Z: you need an USBDLM.INI like this:

[DriveLetter10]
DriveType=fixed
Letter=I

[DriveLetter20]
Letters=U-Z


Uwe



Nefrit wrote:
> I've got a SQL server that uses an external HDD to do backup/transaction
> logs. The drive specifically has to be the "I" drive in order for the
> scripts to run.
>
> All was working fine untill I replaced the 5 external drives with new
> ones. Same size, same brand, different enclosure.
>
> Now whenever I plug the drives in, they do not get an assigned drive
> letter. The only way the server will see them is if I go into "Manage
> your computer" and then assign the drive the proper letter.
>
> This is one of those: I've never seen that before.
>
> Do you guys have any ideas?
> _________________
> 'Home Mortgage Refinancing'
> (http://www.mortgagerefinancing-central.com) 'Make up Games'
> (http://www.dressupwho.com/categories/5/makeovers.html)
>
>
>
>
 
"Uwe Sieber" <mail@uwe-sieber.de> wrote in message
news:5rnu3cF15du7lU1@mid.individual.net...
>
> Identical USB hardware serial are a violation of the
> USB specs. Two of these drive will not work at the
> same time offhand.
>

Although that was the intention, the specification was, in reality, badly
written. You will find that identical serial numbers can be interpreted as
not being a violation of the specification as it is actually worded. This
has caused many people a lot of problems.
 
M.I.5¾ wrote:
> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
> news:5rnu3cF15du7lU1@mid.individual.net...
>> Identical USB hardware serial are a violation of the
>> USB specs. Two of these drive will not work at the
>> same time offhand.
>>
>
> Although that was the intention, the specification was, in reality, badly
> written. You will find that identical serial numbers can be interpreted as
> not being a violation of the specification as it is actually worded. This
> has caused many people a lot of problems.

The USB mass storage specs from June 2003 say
"If provided, the serial number must be unique to each USB
Vendor ID and Product ID pair."
http://www.usb.org/developers/devclass_docs/usb_msc_cbi_1.1.pdf

I don't know the elder revisions of this document but 2003
should be old enough to build correct devices today :-)


Uwe
 
"Uwe Sieber" <mail@uwe-sieber.de> wrote in message
news:5rq2hbF166rdpU1@mid.individual.net...
> M.I.5¾ wrote:
>> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
>> news:5rnu3cF15du7lU1@mid.individual.net...
>>> Identical USB hardware serial are a violation of the
>>> USB specs. Two of these drive will not work at the
>>> same time offhand.
>>>
>>
>> Although that was the intention, the specification was, in reality, badly
>> written. You will find that identical serial numbers can be interpreted
>> as not being a violation of the specification as it is actually worded.
>> This has caused many people a lot of problems.
>
> The USB mass storage specs from June 2003 say
> "If provided, the serial number must be unique to each USB
> Vendor ID and Product ID pair."
> http://www.usb.org/developers/devclass_docs/usb_msc_cbi_1.1.pdf
>
> I don't know the elder revisions of this document but 2003
> should be old enough to build correct devices today :-)
>

I am aware of the spec wording, and it is certainly not at all clear. Your
interpretation is but one interpretation, and presumably makes sense to you
in the light of your knowledge. Unfortunately it has been frequently
interpreted as allowing the *same* devices from the same manufacturer to
have the *same* serial number, albeit *unique* to that product line from
that manufacturer (that is unique to the vendor/product ID pair as
required). This is a valid interpretation of the words as written.

Your interpretation was indeed the intention, but as ever, manufacturers
will happily put an alternative interpretation if it saves them money. It
is by no means the only ambiguous requirement that has cropped up in the
various USB specifications.
 
M.I.5¾ wrote:
> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
> news:5rq2hbF166rdpU1@mid.individual.net...
>> M.I.5¾ wrote:
>>> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
>>> news:5rnu3cF15du7lU1@mid.individual.net...
>>>> Identical USB hardware serial are a violation of the
>>>> USB specs. Two of these drive will not work at the
>>>> same time offhand.
>>>>
>>> Although that was the intention, the specification was, in reality, badly
>>> written. You will find that identical serial numbers can be interpreted
>>> as not being a violation of the specification as it is actually worded.
>>> This has caused many people a lot of problems.
>> The USB mass storage specs from June 2003 say
>> "If provided, the serial number must be unique to each USB
>> Vendor ID and Product ID pair."
>> http://www.usb.org/developers/devclass_docs/usb_msc_cbi_1.1.pdf
>>
>> I don't know the elder revisions of this document but 2003
>> should be old enough to build correct devices today :-)
>>
>
> I am aware of the spec wording, and it is certainly not at all clear. Your
> interpretation is but one interpretation, and presumably makes sense to you
> in the light of your knowledge. Unfortunately it has been frequently
> interpreted as allowing the *same* devices from the same manufacturer to
> have the *same* serial number, albeit *unique* to that product line from
> that manufacturer (that is unique to the vendor/product ID pair as
> required). This is a valid interpretation of the words as written.
>
> Your interpretation was indeed the intention, but as ever, manufacturers
> will happily put an alternative interpretation if it saves them money. It
> is by no means the only ambiguous requirement that has cropped up in the
> various USB specifications.


Yes, now I see that 'all the same serial' is a valid interpretation
of the words as written. It just seemed to be absolutely obvious
to me that a non unique serial is useless, so the other interpretation
never came in my mind...


Uwe
 
"Uwe Sieber" <mail@uwe-sieber.de> wrote in message
news:5rs9vrF15ojvpU1@mid.individual.net...
> M.I.5¾ wrote:
>> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
>> news:5rq2hbF166rdpU1@mid.individual.net...
>>> M.I.5¾ wrote:
>>>> "Uwe Sieber" <mail@uwe-sieber.de> wrote in message
>>>> news:5rnu3cF15du7lU1@mid.individual.net...
>>>>> Identical USB hardware serial are a violation of the
>>>>> USB specs. Two of these drive will not work at the
>>>>> same time offhand.
>>>>>
>>>> Although that was the intention, the specification was, in reality,
>>>> badly written. You will find that identical serial numbers can be
>>>> interpreted as not being a violation of the specification as it is
>>>> actually worded. This has caused many people a lot of problems.
>>> The USB mass storage specs from June 2003 say
>>> "If provided, the serial number must be unique to each USB
>>> Vendor ID and Product ID pair."
>>> http://www.usb.org/developers/devclass_docs/usb_msc_cbi_1.1.pdf
>>>
>>> I don't know the elder revisions of this document but 2003
>>> should be old enough to build correct devices today :-)
>>>
>>
>> I am aware of the spec wording, and it is certainly not at all clear.
>> Your interpretation is but one interpretation, and presumably makes sense
>> to you in the light of your knowledge. Unfortunately it has been
>> frequently interpreted as allowing the *same* devices from the same
>> manufacturer to have the *same* serial number, albeit *unique* to that
>> product line from that manufacturer (that is unique to the vendor/product
>> ID pair as required). This is a valid interpretation of the words as
>> written.
>>
>> Your interpretation was indeed the intention, but as ever, manufacturers
>> will happily put an alternative interpretation if it saves them money.
>> It is by no means the only ambiguous requirement that has cropped up in
>> the various USB specifications.
>
>
> Yes, now I see that 'all the same serial' is a valid interpretation
> of the words as written. It just seemed to be absolutely obvious
> to me that a non unique serial is useless, so the other interpretation
> never came in my mind...
>

But you weren't thinking like a manufacturer, wanting to pinch a penny or
two off his manufacturing costs, and still be able to describe his products
as compliant.

A similar thing happened when the USB2 specification came out covering the
480 Mbps mode of operation. Again, because of an ambiguity, many
manufacturers noticed that the ability to support 480 Mbps operation could
be interpreted as not being a specific requirement of the spec. Thus PCs
and peripherals appeared that were advertised as USB2 compliant, that were,
in fact, unchanged from USB1.1.
 
Back
Top