f



Do IDE RAID controllers exist?

I'm quite familiar with SCSI RAID controllers, but given the large
capacities in modern IDE drives, I was thinking about potentially buying
an IDE RAID controller, but from what I've found in the kernel as far as
support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
devices -> Support for IDE Raid controllers), it appears to be software
RAID only.

As I'm looking to deploy this on a very busy web/database server, I'd
rather not use software RAID, but am I wrong in my assumption then that
hardware IDE RAID controllers don't exist?

Cheers,

Mark
0
10/6/2003 8:03:10 PM
comp.os.linux.hardware 9621 articles. 0 followers. Post Follow

16 Replies
337 Views

Similar Articles

[PageSpeed] 5

Mark Kaufer <mark.kaufer@no.spam.bloodyhell.us> wrote:
> 
> I'm quite familiar with SCSI RAID controllers, but given the large
> capacities in modern IDE drives, I was thinking about potentially buying
> an IDE RAID controller, but from what I've found in the kernel as far as
> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
> devices -> Support for IDE Raid controllers), it appears to be software
> RAID only.
> 
> As I'm looking to deploy this on a very busy web/database server, I'd
> rather not use software RAID, but am I wrong in my assumption then that
> hardware IDE RAID controllers don't exist?

     Yes, hardware IDE RAID controllers exist, 3Ware makes them, Adaptec 
has made them (don't know if they currently are or not).  The 3Ware 
controllers have drivers in the current kernels, and are treated as if 
they were SCSI controllers.

				Scott

0
sdo
10/6/2003 8:13:44 PM
sdo(at)NOSPAM.visi.com writes:

>> I'm quite familiar with SCSI RAID controllers, but given the large
>> capacities in modern IDE drives, I was thinking about potentially buying
>> an IDE RAID controller, but from what I've found in the kernel as far as
>> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
>> devices -> Support for IDE Raid controllers), it appears to be software
>> RAID only.
>> 
>> As I'm looking to deploy this on a very busy web/database server, I'd
>> rather not use software RAID, but am I wrong in my assumption then that
>> hardware IDE RAID controllers don't exist?
>
>      Yes, hardware IDE RAID controllers exist, 3Ware makes them, Adaptec 
> has made them (don't know if they currently are or not).  The 3Ware 
> controllers have drivers in the current kernels, and are treated as if 
> they were SCSI controllers.

Adaptec makes them.  They're cards also appear as SCSI devices,
drivers are in kernels.  Be careful if shopping for Adaptec cards, as
some of them are software RAID.  The price tag usually lets you tell
the difference.

-- 
M�ns Rullg�rd
mru@users.sf.net
0
mru (71)
10/6/2003 8:37:55 PM
After a long battle with technology,Mark Kaufer <mark.kaufer@NO.SPAM.bloodyhell.us>, an earthling, wrote:
> I'm quite familiar with SCSI RAID controllers, but given the large
> capacities in modern IDE drives, I was thinking about potentially buying
> an IDE RAID controller, but from what I've found in the kernel as far as
> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
> devices -> Support for IDE Raid controllers), it appears to be software
> RAID only.
>
> As I'm looking to deploy this on a very busy web/database server, I'd
> rather not use software RAID, but am I wrong in my assumption then that
> hardware IDE RAID controllers don't exist?

You are wrong in that assumption.

3Ware sells some that are fairly well regarded.  Promise sells some
that are slightly less well regarded.  LSI Logic sells some pretty
good ones too.
-- 
If this was helpful, <http://svcs.affero.net/rm.php?r=cbbrowne> rate me
http://www.ntlug.org/~cbbrowne/unix.html
"The true  measure of a  man is how he treats  someone who can  do him
absolutely no good." -- Samuel Johnson, lexicographer (1709-1784)
0
cbbrowne (1108)
10/6/2003 8:42:14 PM
In the last exciting episode, mru@users.sourceforge.net (M�ns Rullg�rd) wrote:
> sdo(at)NOSPAM.visi.com writes:
>
>>> I'm quite familiar with SCSI RAID controllers, but given the large
>>> capacities in modern IDE drives, I was thinking about potentially buying
>>> an IDE RAID controller, but from what I've found in the kernel as far as
>>> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
>>> devices -> Support for IDE Raid controllers), it appears to be software
>>> RAID only.
>>> 
>>> As I'm looking to deploy this on a very busy web/database server, I'd
>>> rather not use software RAID, but am I wrong in my assumption then that
>>> hardware IDE RAID controllers don't exist?
>>
>>      Yes, hardware IDE RAID controllers exist, 3Ware makes them, Adaptec 
>> has made them (don't know if they currently are or not).  The 3Ware 
>> controllers have drivers in the current kernels, and are treated as if 
>> they were SCSI controllers.
>
> Adaptec makes them.  They're cards also appear as SCSI devices,
> drivers are in kernels.  Be careful if shopping for Adaptec cards, as
> some of them are software RAID.  The price tag usually lets you tell
> the difference.

For a while, it was sort of looking like Adaptec was going to be the
"only" controller vendor.

But now, they frankly don't have any decent offerings at the high end,
compared to _vastly_ higher performance stuff from LSI Logic, Intel,
and Compaq.  

They have lots of SCSI controllers at the midrange, supporting all the
fast I/O buses, but they haven't a single product with battery backed
cache, which disqualifies them from _my_ interest.  If I'm building a
big SCSI bus, I want battery backed cache, and Adaptec doesn't have
any controllers that support that.
-- 
(format nil "~S@~S" "aa454" "freenet.carleton.ca")
http://www.ntlug.org/~cbbrowne/oses.html
"So, when you typed in the date, it exploded into a sheet of blue
flame and burned the entire admin wing to the ground? Yes, that's a
known bug. We'll be fixing it in the next release. Until then, try not
to use European date format, and keep an extinguisher handy."
-- slam@pobox.com (Tequila Rapide) 
0
cbbrowne (1108)
10/6/2003 10:07:04 PM
On 6 Oct 2003 22:07:04 GMT, Christopher Browne <cbbrowne@acm.org> wrote:
> and Compaq.  
> 
> They have lots of SCSI controllers at the midrange, supporting all the
> fast I/O buses, but they haven't a single product with battery backed
> cache, which disqualifies them from _my_ interest.  If I'm building a
> big SCSI bus, I want battery backed cache, and Adaptec doesn't have
> any controllers that support that.

If you want battery backup, use a UPS.

Are you also bothered by the fact that no processor has battery backed
cache?
0
TCS
10/6/2003 10:34:39 PM
Oops! TCS <The.Central.Scrutinizer@p.o.b.o.x.com> was seen spray-painting on a wall:
> On 6 Oct 2003 22:07:04 GMT, Christopher Browne <cbbrowne@acm.org> wrote:
>> and Compaq.  
>> 
>> They have lots of SCSI controllers at the midrange, supporting all the
>> fast I/O buses, but they haven't a single product with battery backed
>> cache, which disqualifies them from _my_ interest.  If I'm building a
>> big SCSI bus, I want battery backed cache, and Adaptec doesn't have
>> any controllers that support that.
>
> If you want battery backup, use a UPS.

I already do.  We just upgraded the UPS at the office; it now has
three units, each consisting of about 40 marine-class gel cells.

> Are you also bothered by the fact that no processor has battery backed
> cache?

No, that doesn't trouble me.

If the CPU goes down, then half-baked database transactions will roll
back without any particular problems or confusion.

If, on the other hand, the system claims that particular transactions
are committed, because they have made it into the RAID controller's
cache, I want that to be the case even if power goes down, whether on
the local grid, local block, or the computer's power supply.

Battery backed cache can give, _without loss of safety_, on the order
of a 20x speedup for OLTP systems.
-- 
select 'aa454' || '@' || 'freenet.carleton.ca';
http://www3.sympatico.ca/cbbrowne/x.html
"My mom said she learned how to swim. Someone took her out in the lake
and threw  her off  the boat. That's  how she  learned how to  swim. I
said, 'Mom, they  weren't trying to teach you how  to swim.' " 
-- Paula Poundstone
0
cbbrowne (1108)
10/7/2003 2:11:53 AM
Christopher Browne <cbbrowne@acm.org> wrote:
>> If you want battery backup, use a UPS.
> 
> I already do.  We just upgraded the UPS at the office; it now has
> three units, each consisting of about 40 marine-class gel cells.
> 
>> Are you also bothered by the fact that no processor has battery backed
>> cache?
> 
> No, that doesn't trouble me.
> 
> If the CPU goes down, then half-baked database transactions will roll
> back without any particular problems or confusion.
> 
> If, on the other hand, the system claims that particular transactions
> are committed, because they have made it into the RAID controller's
> cache, I want that to be the case even if power goes down, whether on
> the local grid, local block, or the computer's power supply.
> 
> Battery backed cache can give, _without loss of safety_, on the order
> of a 20x speedup for OLTP systems.

Unless that battery shorts out.  For OP... what was wrong with software
RAID?

-- 
William Park, Open Geometry Consulting, <opengeometry@yahoo.ca>
Linux solution for data management and processing. 
0
opengeometry (485)
10/7/2003 4:35:42 AM
On 6 Oct 2003 20:42:14 GMT, Christopher Browne <cbbrowne@acm.org> wrote to comp.os.linux.hardware:
> After a long battle with technology,Mark Kaufer <mark.kaufer@NO.SPAM.bloodyhell.us>, an earthling, wrote:
>> I'm quite familiar with SCSI RAID controllers, but given the large
>> capacities in modern IDE drives, I was thinking about potentially buying
>> an IDE RAID controller, but from what I've found in the kernel as far as
>> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
>> devices -> Support for IDE Raid controllers), it appears to be software
>> RAID only.
>>
>> As I'm looking to deploy this on a very busy web/database server, I'd
>> rather not use software RAID, but am I wrong in my assumption then that
>> hardware IDE RAID controllers don't exist?
> 
> You are wrong in that assumption.
> 
> 3Ware sells some that are fairly well regarded.  Promise sells some
> that are slightly less well regarded.  LSI Logic sells some pretty
> good ones too.

Thanks to all who replied.

Please forgive my ignorance in the subject, but are we talking serial
ATA (Escalade 8506) or parallel ATA (Escalade 7506) ?  I've found the
support for these 3Ware devices in the kernel (3W_XXXX_RAID) but it
isn't exactly clear which of these the support is for.

Thanks again,

Mark
0
10/7/2003 5:03:12 AM
Mark Kaufer <mark.kaufer@no.spam.bloodyhell.us> wrote:
>>> I'm quite familiar with SCSI RAID controllers, but given the large
>>> capacities in modern IDE drives, I was thinking about potentially buying
>>> an IDE RAID controller, but from what I've found in the kernel as far as
>>> support goes, (ATA/IDE/MFM/RLL support -> IDE, ATA and ATAPI Block
>>> devices -> Support for IDE Raid controllers), it appears to be software
>>> RAID only.
>>>
>>> As I'm looking to deploy this on a very busy web/database server, I'd
>>> rather not use software RAID, but am I wrong in my assumption then that
>>> hardware IDE RAID controllers don't exist?
>> 
>> You are wrong in that assumption.
>> 
>> 3Ware sells some that are fairly well regarded.  Promise sells some
>> that are slightly less well regarded.  LSI Logic sells some pretty
>> good ones too.
> 
> Thanks to all who replied.
> 
> Please forgive my ignorance in the subject, but are we talking serial
> ATA (Escalade 8506) or parallel ATA (Escalade 7506) ?  I've found the
> support for these 3Ware devices in the kernel (3W_XXXX_RAID) but it
> isn't exactly clear which of these the support is for.

    Both.  The parallel ATA boards have been around a good deal longer, 
but they use the same drivers at the kernel level.

				Scott
0
sdo
10/7/2003 12:16:31 PM
> If you want battery backup, use a UPS.
>
> Are you also bothered by the fact that no processor has battery backed
> cache?

  A UPS buys you absolutely nothing when a technician/janitor/whoever
kicks or pulls the cord out of the back of the machine.

steve


0
unt (107)
10/7/2003 4:50:44 PM
On 07 Oct 2003 12:16:31 GMT, sdo(at)NOSPAM.visi.com <> wrote to comp.os.linux.hardware:
> Mark Kaufer <mark.kaufer@no.spam.bloodyhell.us> wrote:
> 
>     Both.  The parallel ATA boards have been around a good deal longer, 
> but they use the same drivers at the kernel level.

Thank you all for the replies and the insight.  This has been very
helpful.

Best regards,

Mark
0
10/7/2003 10:39:37 PM
On Tue, 7 Oct 2003 10:50:44 -0600, Steve Wolfe <unt@see.signature.com> wrote:
>> If you want battery backup, use a UPS.
>>
>> Are you also bothered by the fact that no processor has battery backed
>> cache?

>  A UPS buys you absolutely nothing when a technician/janitor/whoever
>kicks or pulls the cord out of the back of the machine.

A battery backed cache buys you even less.
0
TCS
10/8/2003 2:55:58 AM
TCS <The.Central.Scrutinizer@p.o.b.o.x.com> wrote:
> On Tue, 7 Oct 2003 10:50:44 -0600, Steve Wolfe <unt@see.signature.com>
> wrote:
>>> If you want battery backup, use a UPS.
>>>
>>> Are you also bothered by the fact that no processor has battery
>>> backed cache?
> 
>>  A UPS buys you absolutely nothing when a technician/janitor/whoever
>>kicks or pulls the cord out of the back of the machine.
> 
> A battery backed cache buys you even less.

True.  But, you can certainly brag about it to "IT manager"...

-- 
William Park, Open Geometry Consulting, <opengeometry@yahoo.ca>
Linux solution for data management and processing. 
0
opengeometry (485)
10/8/2003 5:38:43 AM
After takin a swig o' Arrakan spice grog, TCS <The.Central.Scrutinizer@p.o.b.o.x.com> belched out...:
> On Tue, 7 Oct 2003 10:50:44 -0600, Steve Wolfe <unt@see.signature.com> wrote:
>>> If you want battery backup, use a UPS.
>>>
>>> Are you also bothered by the fact that no processor has battery backed
>>> cache?
>
>>  A UPS buys you absolutely nothing when a technician/janitor/whoever
>>kicks or pulls the cord out of the back of the machine.
>
> A battery backed cache buys you even less.

Untrue.  It means that transactions can commit properly when power is
restored.
-- 
"cbbrowne","@","acm.org"
http://www3.sympatico.ca/cbbrowne/emacs.html
"I will not send lard through the mail" ^ 100 -- Bart Simpson
0
cbbrowne (1108)
10/8/2003 2:02:28 PM
On 8 Oct 2003 14:02:28 GMT, Christopher Browne <cbbrowne@acm.org> wrote:
>After takin a swig o' Arrakan spice grog, TCS <The.Central.Scrutinizer@p.o.b.o.x.com> belched out...:
>> On Tue, 7 Oct 2003 10:50:44 -0600, Steve Wolfe <unt@see.signature.com> wrote:
>>>> If you want battery backup, use a UPS.
>>>>
>>>> Are you also bothered by the fact that no processor has battery backed
>>>> cache?
>>
>>>  A UPS buys you absolutely nothing when a technician/janitor/whoever
>>>kicks or pulls the cord out of the back of the machine.
>>
>> A battery backed cache buys you even less.

>Untrue.  It means that transactions can commit properly when power is
>restored.

No.  It means the cache is nonvoltile while the drive is spinning down
and offline and the CPU is crashing and sending randome junk to the
controller.  worthless.
0
TCS
10/9/2003 2:17:57 AM
> >> Are you also bothered by the fact that no processor has battery backed
> >> cache?
>
> >  A UPS buys you absolutely nothing when a technician/janitor/whoever
> >kicks or pulls the cord out of the back of the machine.
>
> A battery backed cache buys you even less.

  I don't know, the idea of the RAID controller booting up, seeing the data
still in cache, and flushing it to disk as if there had been no interruption
sounds awfully nice to me.

steve



0
unt1 (176)
10/9/2003 4:02:53 AM
Reply: