It turns out the server we were going to acquire has 2.5" SAS drives.
I'd never heard of same and assumed the slots/sleds were the usual 3.5"
size; the kind where you can use the {orders-of-magnitude..} cheaper and
4x larger SATA drives transparently.
Looking around, it appears they are speciality drives, 10,000 or
15,000 RPM. I do not see any of them larger than 140 GB; most are
smaller. Further, Hitachi stopped making them two years ago, and it's
mostly a Seagate concept.
Anyone know the history on them?
--
A host is a host from coast to coast.................wb8foz@nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433
David Lesher wrote:
>
> I've just tumbled into a situation re: Subject:
>
> It turns out the server we were going to acquire has 2.5" SAS drives.
> I'd never heard of same and assumed the slots/sleds were the usual 3.5"
> size; the kind where you can use the {orders-of-magnitude..} cheaper and
> 4x larger SATA drives transparently.
>
> Looking around, it appears they are speciality drives, 10,000 or
> 15,000 RPM. I do not see any of them larger than 140 GB; most are
> smaller. Further, Hitachi stopped making them two years ago, and it's
> mostly a Seagate concept.
>
> Anyone know the history on them?
IIRC the 2.5" server disks are introduced by Seagate with the Savvio
series. Because of the desired fast access times of server disks, some
of them use media with smaller diameter even if the case is 3.5" (to
reduce the distance the heads have to move). Look at this picture of a
3.5" Seagate Cheetah X15 for an example: http://images.ciao.com/ide/images/pr...ct-1240879.jpg
This is one of the reasons why 3.5" server disks have smaller capacities
compared to consumer ones with full-size media. Seagate now thought that
it should be possible to put such a disk into a 2.5" case without losing
too much performance (the goal was to save space in disk arrays). The
first such drives had a parallel SCSI interface and a SCA2 connector.
Now they migrate to SAS.
"David Lesher" <wb8foz@panix.com> wrote in message news:fkju66$r1$1@reader1.panix.com...
> I've just tumbled into a situation re: Subject:
>
> It turns out the server we were going to acquire has 2.5" SAS drives.
> I'd never heard of same and assumed the slots/sleds were the usual 3.5"
> size; the kind where you can use the {orders-of-magnitude..} cheaper and
> 4x larger SATA drives transparently.
>
> Looking around, it appears they are speciality drives, 10,000 or
> 15,000 RPM. I do not see any of them larger than 140 GB; most are
> smaller. Further, Hitachi stopped making them two years ago, and it's
> mostly a Seagate concept.
>
> Anyone know the history on them?
>
This certainly is one of the industries stupidest ideas.
A 10K 3.5" drive has 3" platters, the 2.5" platters have half the capacity.
3.5" drives come with 1-5 platters, 2.5" only 1-2. So
total capacity is 1/4 of 10K or 1/2 of 15K 3.5" drives,
but cost for SCSI drives has alway been per platter (or head).
If lowest $/GB is your goal, 7200 SATA will be even cheaper and cooler.
You don't get higher density (GB per frontal area),
you don't get lower power over 4-5 platter 10K drives,
the only benefit is more IO/s because you have more actuators.
In comp.periphs.scsi Eric Gisin <gisin@uniserve.com> wrote:
> "David Lesher" <wb8foz@panix.com> wrote in message news:fkju66$r1$1@reader1.panix.com...
>> I've just tumbled into a situation re: Subject:
>>
>> It turns out the server we were going to acquire has 2.5" SAS drives.
>> I'd never heard of same and assumed the slots/sleds were the usual 3.5"
>> size; the kind where you can use the {orders-of-magnitude..} cheaper and
>> 4x larger SATA drives transparently.
>>
>> Looking around, it appears they are speciality drives, 10,000 or
>> 15,000 RPM. I do not see any of them larger than 140 GB; most are
>> smaller. Further, Hitachi stopped making them two years ago, and it's
>> mostly a Seagate concept.
>>
>> Anyone know the history on them?
>>
> This certainly is one of the industries stupidest ideas.
>
> A 10K 3.5" drive has 3" platters, the 2.5" platters have half the capacity.
> 3.5" drives come with 1-5 platters, 2.5" only 1-2. So
> total capacity is 1/4 of 10K or 1/2 of 15K 3.5" drives,
> but cost for SCSI drives has alway been per platter (or head).
> If lowest $/GB is your goal, 7200 SATA will be even cheaper and cooler.
>
> You don't get higher density (GB per frontal area),
> you don't get lower power over 4-5 platter 10K drives,
> the only benefit is more IO/s because you have more actuators.
Eric Gisin wrote in news:13mt56fs4061g20@corp.supernews.com
> "David Lesher" <wb8foz@panix.com> wrote in message news:fkju66$r1$1@reader1.panix.com...
> > I've just tumbled into a situation re: Subject:
> >
> > It turns out the server we were going to acquire has 2.5" SAS drives.
> > I'd never heard of same and assumed the slots/sleds were the usual 3.5"
> > size; the kind where you can use the {orders-of-magnitude..} cheaper and
> > 4x larger SATA drives transparently.
> >
> > Looking around, it appears they are speciality drives, 10,000 or
> > 15,000 RPM. I do not see any of them larger than 140 GB; most are
> > smaller. Further, Hitachi stopped making them two years ago, and it's
> > mostly a Seagate concept.
> >
> > Anyone know the history on them?
> >
> This certainly is one of the industries stupidest ideas.
> A 10K 3.5" drive has 3" platters, the 2.5" platters have half the capacity.
Pity 2.5" platters won't fit in a 2.5" drive.
> 3.5" drives come with 1-5 platters, 2.5" only 1-2.
> So total capacity is 1/4 of 10K or 1/2 of 15K 3.5" drives,
> but cost for SCSI drives has alway been per platter (or head).
> If lowest $/GB is your goal, 7200 SATA will be even cheaper and cooler.
> You don't get higher density (GB per frontal area),
> you don't get lower power over 4-5 platter 10K drives,
> the only benefit is more IO/s because you have more actuators.
David Lesher wrote in news:fkju66$r1$1@reader1.panix.com
> I've just tumbled into a situation re: Subject:
> It turns out the server we were going to acquire has 2.5" SAS drives.
Just a minor detail.
> I'd never heard of same and assumed the slots/sleds were the usual 3.5"
> size;
Yeah, why mention 2.5" when it's actually 3.5".
> the kind where you can use the {orders-of-magnitude..} cheaper and
> 4x larger SATA drives transparently.
>
> Looking around, it appears they are speciality drives, 10,000 or 15,000
> RPM. I do not see any of them larger than 140 GB; most are smaller.
> Further, Hitachi stopped making them two years ago,
Really?
They stopped making them already before started making them. Clever!
Eric Gisin wrote:
> "David Lesher" <wb8foz@panix.com> wrote in message
> news:fkju66$r1$1@reader1.panix.com...
>>
>> Looking around, it appears they are speciality drives, 10,000 or
>> 15,000 RPM. I do not see any of them larger than 140 GB; most are
>> smaller. Further, Hitachi stopped making them two years ago, and it's
>> mostly a Seagate concept.
>>
>> Anyone know the history on them?
>>
> This certainly is one of the industries stupidest ideas.
A lot of people would noot agree.
> A 10K 3.5" drive has 3" platters, the 2.5" platters have half the capacity.
> 3.5" drives come with 1-5 platters, 2.5" only 1-2. So
> total capacity is 1/4 of 10K or 1/2 of 15K 3.5" drives, but cost for
> SCSI drives has alway been per platter (or head).
> If lowest $/GB is your goal, 7200 SATA will be even cheaper and cooler.
>
> You don't get higher density (GB per frontal area), you don't get lower
> power over 4-5 platter 10K drives,
> the only benefit is more IO/s because you have more actuators.
This benefit is more important for business use than pure capacity.
I don't like 2,5" driver either, but for blade servers / virtualization
and important data on SAN/NAS or something else, 2,5" disks are good enough.
In article <13mt56fs4061g20@corp.supernews.com>,
Eric Gisin <gisin@uniserve.com> wrote:
>
>You don't get higher density (GB per frontal area),
>you don't get lower power over 4-5 platter 10K drives,
>the only benefit is more IO/s because you have more actuators.
I'll take more I/O per second in the same space over almost any
other option, thanks.
>>You don't get higher density (GB per frontal area),
>>you don't get lower power over 4-5 platter 10K drives,
>>the only benefit is more IO/s because you have more actuators.
>I'll take more I/O per second in the same space over almost any
>other option, thanks.
Depends on what you are after. We need 650GB for archival storage. We can
buy 2 {RAID, you know} 750GB at ~$200 each. Or we can buy 10 of the 2.5"
160GB's at $250 each. [We'll also need some way to mount the six that
won't fit in the server.]
Given the usage of the server, which is a tortoise vs hare case; which
should we do?
--
A host is a host from coast to coast.................wb8foz@nrk.com
& no one will talk to a host that's close........[v].(301) 56-LINUX
Unless the host (that isn't close).........................pob 1433
is busy, hung or dead....................................20915-1433
David Lesher wrote:
> tls@panix.com (Thor Lancelot Simon) writes:
>
>> I'll take more I/O per second in the same space over almost any
>> other option, thanks.
>
> Depends on what you are after. We need 650GB for archival storage. We can
> buy 2 {RAID, you know} 750GB at ~$200 each. Or we can buy 10 of the 2.5"
> 160GB's at $250 each. [We'll also need some way to mount the six that
> won't fit in the server.]
In that case, you define what you need. In most business needs speed is
essential, not pure capacity.
> Given the usage of the server, which is a tortoise vs hare case; which
> should we do?
Exactly what you did. But it doesn't have to be the best case scenario
for others.
2.5" SAS drives 
Linear Mode
