how to know a hard drive will mix well

how to know a hard drive will mix well

[David T-G]

Hi, all --

I have

  diskfarm:~ # mdadm -D /dev/md51 | egrep /dev/
  /dev/md51:
	 0     259        9        0      active sync   /dev/sdb51
	 1     259        2        1      active sync   /dev/sdc51
	 3     259       16        2      active sync   /dev/sdd51
	 4     259       23        3      active sync   /dev/sdj51

Toshiba X300 10T

  diskfarm:~ # smartctl -i /dev/sdb | egrep '(Model|Number|Version):'
  Device Model:     TOSHIBA HDWR11A
  Serial Number:    61U0A0HQFBKG
  Firmware Version: 0603

drives in my disk farm, and it's about time to grow again.  As I
shopped around, I stumbled over a WD Red Plus 12T WD120EFBX drive at
just $10 more, and who wouldn't want an extra 2T for ten bucks?  So I'm
contemplating rolling in a different model.  But how do I know that it
will fit well into the mix?

The X300 is a 7200rpm 256Mcache 6G/sec CMR drive.  The Red Plus is also
listed as a 7200rpm* 256Mcache 6G/sec CMR drive.  They certainly sound
equivalent.  What else do I need to consider, and where else do I need
to look to learn?

* The spec actually says "7200 RPM Class".  Does that mean not really
7200rpm?  That wouldn't surprise me in this modern day and age, and if
the X300 also isn't really then that also wouldn't.


TIA & HAND

:-D
-- 
David T-G
See http://justpickone.org/davidtg/email/
See http://justpickone.org/davidtg/tofu.txt

Re: how to know a hard drive will mix well

[Wols Lists]

On 02/02/2023 12:43, David T-G wrote:
> Hi, all --
> 
> I have
> 
>    diskfarm:~ # mdadm -D /dev/md51 | egrep /dev/
>    /dev/md51:
> 	 0     259        9        0      active sync   /dev/sdb51
> 	 1     259        2        1      active sync   /dev/sdc51
> 	 3     259       16        2      active sync   /dev/sdd51
> 	 4     259       23        3      active sync   /dev/sdj51
> 
> Toshiba X300 10T
> 
>    diskfarm:~ # smartctl -i /dev/sdb | egrep '(Model|Number|Version):'
>    Device Model:     TOSHIBA HDWR11A
>    Serial Number:    61U0A0HQFBKG
>    Firmware Version: 0603
> 
> drives in my disk farm, and it's about time to grow again.  As I
> shopped around, I stumbled over a WD Red Plus 12T WD120EFBX drive at
> just $10 more, and who wouldn't want an extra 2T for ten bucks?  So I'm
> contemplating rolling in a different model.  But how do I know that it
> will fit well into the mix?
> 
> The X300 is a 7200rpm 256Mcache 6G/sec CMR drive.  The Red Plus is also
> listed as a 7200rpm* 256Mcache 6G/sec CMR drive.  They certainly sound
> equivalent.  What else do I need to consider, and where else do I need
> to look to learn?

I don't trust WD. That said, a lot of people do. I've got two 4TB 
Ironwolves, one 3TB Barracuda (slap wrist!), and one 8TB X300.

So mixing drives isn't a problem - just take a look at the Ironwolf - 
that might give you extra capacity too.
> 
> * The spec actually says "7200 RPM Class".  Does that mean not really
> 7200rpm?  That wouldn't surprise me in this modern day and age, and if
> the X300 also isn't really then that also wouldn't.
> 
> 
Drives now are "Constant Head Speed" not constant rpm. I'm guessing, 
that what it means is that for the inner tracks it spins at 7200, and as 
the heads move out, the rpms slow down to keep the speed the head is 
going over the platter constant.

Cheers,
Wol

Re: how to know a hard drive will mix well

[Pascal Hambourg]

On 02/02/2023 at 19:52, Wols Lists wrote:
>
> Drives now are "Constant Head Speed" not constant rpm. I'm guessing, 

Source ?

"Constant linear velocity" (CLV, as opposed to "constant angular 
velocity" - CAV) is adapted for use requiring a constant transfer rate 
such as media recording or playback, not random access, and is mostly 
found in optical disc drives.

> that what it means is that for the inner tracks it spins at 7200, and as 
> the heads move out, the rpms slow down to keep the speed the head is 
> going over the platter constant.

It also implies that the spindle is able to accelerate from the lower 
speed (around 3600 RPM) to 7200 RPM or decelerate from 7200 to 3600 RPM 
in less than the full stroke time, which is around 20 ms in order to not 
degrade the access time. It sounds highly unlikely to me.

Re: how to know a hard drive will mix well

[Wol]

On 02/02/2023 22:47, Pascal Hambourg wrote:
> On 02/02/2023 at 19:52, Wols Lists wrote:
>>
>> Drives now are "Constant Head Speed" not constant rpm. I'm guessing, 
> 
> Source ?
> 
> "Constant linear velocity" (CLV, as opposed to "constant angular 
> velocity" - CAV) is adapted for use requiring a constant transfer rate 
> such as media recording or playback, not random access, and is mostly 
> found in optical disc drives.
> 
>> that what it means is that for the inner tracks it spins at 7200, and 
>> as the heads move out, the rpms slow down to keep the speed the head 
>> is going over the platter constant.
> 
> It also implies that the spindle is able to accelerate from the lower 
> speed (around 3600 RPM) to 7200 RPM or decelerate from 7200 to 3600 RPM 
> in less than the full stroke time, which is around 20 ms in order to not 
> degrade the access time. It sounds highly unlikely to me.

Just tried to Google it. Did I say I didn't trust WD?

It sounds like it's marketing speak designed to re-define the meaning of 
Revolutions Per Minute.

Apparently 7200rpm drives are being sold as "5400 class". Misleading 
people who are prepared to take the speed hit, into buying a drive that 
is power-hungry and noisy.

As the google page I saw said - "Why describe it as "5400-class" in a 
spec sheet - the people who don't know what it means won't be reading 
spec sheets, and the people who do read spec sheets won't know what the 
hell it means and it is deceptive".

I'd go and look for another X300, or an Ironwolf. At least you'll know 
what you're getting ...

Cheers,
Wol

Re: how to know a hard drive will mix well

[Wol]

Replying to myself,

https://www.amazon.co.uk/Seagate-ST12000VN0008-4TB-ST4000VM000/dp/B07J351T2V/ref=sr_1_5?keywords=12tb+hard+drive&qid=1675380717&sprefix=12TB+has%2Caps%2C77&sr=8-5

https://www.amazon.co.uk/Plus-12TB-Internal-Hard-Drive/dp/B0BNJFHGHR/ref=sr_1_2?crid=MDH53LQ81TE3&keywords=wd+12TB+red+plus&qid=1675380788&sprefix=wd+12tb+red+plus%2Caps%2C64&sr=8-2

The Ironwolf is £50 - 20% - cheaper than the WD. Ditch the WD ...

Cheers,
Wol

On 02/02/2023 23:27, Wol wrote:
> On 02/02/2023 22:47, Pascal Hambourg wrote:
>> On 02/02/2023 at 19:52, Wols Lists wrote:
>>>
>>> Drives now are "Constant Head Speed" not constant rpm. I'm guessing, 
>>
>> Source ?
>>
>> "Constant linear velocity" (CLV, as opposed to "constant angular 
>> velocity" - CAV) is adapted for use requiring a constant transfer rate 
>> such as media recording or playback, not random access, and is mostly 
>> found in optical disc drives.
>>
>>> that what it means is that for the inner tracks it spins at 7200, and 
>>> as the heads move out, the rpms slow down to keep the speed the head 
>>> is going over the platter constant.
>>
>> It also implies that the spindle is able to accelerate from the lower 
>> speed (around 3600 RPM) to 7200 RPM or decelerate from 7200 to 3600 
>> RPM in less than the full stroke time, which is around 20 ms in order 
>> to not degrade the access time. It sounds highly unlikely to me.
> 
> Just tried to Google it. Did I say I didn't trust WD?
> 
> It sounds like it's marketing speak designed to re-define the meaning of 
> Revolutions Per Minute.
> 
> Apparently 7200rpm drives are being sold as "5400 class". Misleading 
> people who are prepared to take the speed hit, into buying a drive that 
> is power-hungry and noisy.
> 
> As the google page I saw said - "Why describe it as "5400-class" in a 
> spec sheet - the people who don't know what it means won't be reading 
> spec sheets, and the people who do read spec sheets won't know what the 
> hell it means and it is deceptive".
> 
> I'd go and look for another X300, or an Ironwolf. At least you'll know 
> what you're getting ...
> 
> Cheers,
> Wol

Re: how to know a hard drive will mix well

[David Thorburn-Gundlach]

Wol, et al --

...and then Wol said...
% 
% Just tried to Google it. Did I say I didn't trust WD?
% 
% It sounds like it's marketing speak designed to re-define the meaning of
% Revolutions Per Minute.

Ew.  Figured it was something like that ...


% 
...
% 
% I'd go and look for another X300, or an Ironwolf. At least you'll know what
% you're getting ...

Thanks.  I've spent quite a bit surfing for Ironwolf, [XN]300, Exos, and
Red drives in various sizes.  Pricing is ... interesting.

Along those lines, does anyone have ACTUAL experience with MaxDigitalData
"white" drives?  I've read that they're everything from OEM bulk unbranded
major-brand-name drives to badly-refurbished used drives, so I don't need
apocrypha, but I have to at least take a look since their prices are about
half the going rate.  I'd love to buy two for the price of one and expand
from RAID5 to RAID6 ...


% 
% Cheers,
% Wol


Thanks again

:-D
-- 
David T-G
See http://justpickone.org/davidtg/email/
See http://justpickone.org/davidtg/tofu.txt

Re: how to know a hard drive will mix well

[David Thorburn-Gundlach]

Roger, et al --

...and then Roger Heflin said...
% I have mixed many drives, I have never really had a major slowness.  At

Good to know.


% worst you will be limited by the drive with the slowest seeks/transfer
% rate.  So long as one disk's RPM is not significantly different you should

That's pretty much what I was wondering, but I imagine if it's the same
RPM and the same cache and the same SATA rate it should be ... about the
same.


% not notice it, unless you are running your array right at the limit.

Hmmph.  I'm not entirely sure I'm not anyway, but it is just a little PC
mobo running primary and daughtercard SATA ports.  A Real Server (tm) is
far, far beyond the current horizon :-)


% 
% Technically you may want to also look at another parm that is the sustained
% transfer rate range coming off the platter.  This will be a range that is
[snip]

Oh, thanks!  Sounds good.


HAND

:-D
-- 
David T-G
See http://justpickone.org/davidtg/email/
See http://justpickone.org/davidtg/tofu.txt

throughput testing (was "Re: how to know a hard drive will mix well")

[David Thorburn-Gundlach]

Roger, et al --

...and then Roger Heflin said...
% On Fri, Feb 3, 2023 at 8:22 PM David Thorburn-Gundlach <
% davidtg-robot@justpickone.org> wrote:
% >
% > % not notice it, unless you are running your array right at the limit.
% >
% > Hmmph.  I'm not entirely sure I'm not anyway, but it is just a little PC
% > mobo running primary and daughtercard SATA ports.  A Real Server (tm) is
% > far, far beyond the current horizon :-)
% 
% It does not take much of a motherboard/cpu to service a bunch of spinning
% disks.
% 
...
% About the only major thing that makes any difference is making sure that
% whatever card is running the sata ports does not have a bottleneck.
[snip]

Thanks!  It'll be a bit before I can play, but I'll do some reading and
work some tests.  Stay tuned :-)


HAND

:-D
-- 
David T-G
See http://justpickone.org/davidtg/email/
See http://justpickone.org/davidtg/tofu.txt