2,5" hdd laptop & Linux

[Don Juan]

I'm looking for a 2,5" sata hdd, for my laptop, and use it with linux distro.
I think to buy a standard device (not a ssd) for a size of 500GB.

I would simply buy a brand-model that hasn't problem with linux.
Does anyone suggest me a name?
I'm worried (after some online search to retrieve info and opinions on models that i've seen on sale mìbranded wd,hitachi..) about: 1) some "bug" in many hdd's firmwares for about the power saving function (sympotms of continuosly sound-clicking cused by continuously parking/reactive of the heads) 2)Others less worried issues as "alignment partition"

https://ata.wiki.kernel.org/index.php/Known_issues#Drives_which_perform_frequent_head_unloads_under_Linux

https://ata.wiki.kernel.org/index.php/ATA_4_KiB_sector_issues

[the wharf rat]

In article <4d535e2f-4491-4c82...@googlegroups.com>,

Don Juan <linuz...@gmail.com> wrote:
>I'm looking for a 2,5" sata hdd, for my laptop, and use it with linux distro.
>I think to buy a standard device (not a ssd) for a size of 500GB.
>
>I would simply buy a brand-model that hasn't problem with linux.
>Does anyone suggest me a name?

The brand of drive won't matter to Linux. The controller might but
almost certainly won't. As far as consumer grade laptop drives go I like
Western Digital or Fujitsu. But there's not that much difference. Samsung
drives are Ok and cheap.

Look into a SSD though. Makes a big performance difference.
Subjectively, won't help benchmarks that much.

[Happy Oyster]

On Sat, 23 Mar 2013 15:06:02 +0000 (UTC), wr...@panix.com (the wharf rat)
wrote:

>In article <4d535e2f-4491-4c82...@googlegroups.com>,
>Don Juan <linuz...@gmail.com> wrote:
>>I'm looking for a 2,5" sata hdd, for my laptop, and use it with linux distro.
>>I think to buy a standard device (not a ssd) for a size of 500GB.
>>
>>I would simply buy a brand-model that hasn't problem with linux.
>>Does anyone suggest me a name?
>
> The brand of drive won't matter to Linux. The controller might but
>almost certainly won't. As far as consumer grade laptop drives go I like
>Western Digital or Fujitsu. But there's not that much difference. Samsung
>drives are Ok and cheap.

Right.

One thing to add: naked HDDs, to be built in internally in a PC, are
quite expensive, compared with external HDDs that have an enclosure and
power supply.

The question is: why build in a LARGE HDD? It only causes trouble. If
the notebook is used mobile, the an extra HDD might be a nuisance, but
if the computer is used in an office only, the HDD being external has
big advantages, may it be security and ease of use.
--

Crowd-founding is for money, crowd-publishing is for mankind.

http://www.supermanpost.com/

[Don Juan]

Il giorno sabato 23 marzo 2013 16:06:02 UTC+1, the wharf rat ha scritto:

> The brand of drive won't matter to Linux. The controller might but
>
> almost certainly won't. As far as consumer grade laptop drives go I like
>
> Western Digital or Fujitsu. But there's not that much difference. Samsung
>
> drives are Ok and cheap.

Ok, but problem is know what model, (now on sale) has these simple features is Linux compatible, .
I've found, for example that latest wd drives are very problematic (issue is head auto parks non standard function) with Linux (and also hitachi or seagate and samsung).
My temporary solution: Find a laptop model sold with Linux and buy their hdd. Actually, with this "heuristc" I've selected a Toshiba MK series that is hdd of hp laptop selled with ubuntu.

> Look into a SSD though. Makes a big performance difference.
>
> Subjectively, won't help benchmarks that much.

Yes the alternative is ssd.

Thanks for reply

[the wharf rat]

In article <0b078ac3-bcc3-4033...@googlegroups.com>,

Don Juan <linuz...@gmail.com> wrote:
>
>Ok, but problem is know what model, (now on sale) has these simple
>features is Linux compatible, .
>I've found, for example that latest wd drives are very problematic
>(issue is head auto parks non standard function) with Linux (and also

The issue is WD's change to 4096 byte blocks. Linux is perfectly
OK with this but Windows is not. WD's fix was to have the drive firmware
lie about the geometry and claim "standard" 512 byte blocks. That breaks
on ANY OS if the partition ends on a boundary not divisible by 8. In Linux,
you can easily work around this nonsense by manually creating partitions
and ensuring this. The drives aren't incompatible.

So don't use the brand new "advanced format" WD drives, use a
real Enterprise drive and SAS, or use an older and cheaper drive for
your Linux system. Or try a Samsung :-)

[Don Juan]

Il giorno martedì 26 marzo 2013 16:30:52 UTC+1, the wharf rat ha scritto:
> >I've found, for example that latest wd drives are very problematic
>
> >(issue is head auto parks non standard function) with Linux (and also
>
>
>
> The issue is WD's change to 4096 byte blocks. Linux is perfectly

Ok, thanks, but I'm worried primarily about this issue:
https://ata.wiki.kernel.org/index.php/Known_issues#Drives_which_perform_frequent_head_unloads_under_Linux

Now on sale there are some hdd with this issue, and solution are trick, or there aren't solution.

I've seen also, in many forum on web, some users speak about failure of these devices, only after a few months. These threads are often about ask suggest for data recovery, but I think this is too frequent and strange, and probably users haven't awareness that with some s.o. their device could have abnormals behaviours with consequnces on devices's life .

However after an attempt with an hdd 2,5", I'm seriously thinking to pass to a ssd device (also because sata-300 on 60/128GB aren't so expensive).
In my netbook hdd slot, is specified compatible with an hdd's highness of 9.5mm. I've seen that numerous interesting ssd are in the 7mm form factor.
Do you think is equal or I will have some trouble with sata connector position, or with holes screw's position (in my netbook, the device is put in a little jail / case with four screws, then slided it until connects it on the sata connector).

[the wharf rat]

In article <66b090bc-56d3-4ace...@googlegroups.com>,

Don Juan <linuz...@gmail.com> wrote:
>
>Ok, thanks, but I'm worried primarily about this issue:
>https://ata.wiki.kernel.org/index.php/Known_issues#Drives_which_perform_frequent_head_unloads_under_Linux
>

Seems pretty much a non-issue. It's not because of Linux (you're
just reading about it on a kernel forum :-), but because of a buggy interaction
between certain laptop IDE drives and power management. It seems to depend
on the installation platform as well as the particular drive since manifestation
is dependent on IO patterns.

The issue here with WD is that WD chooses to require their own
special program to reset APM parameters on a few models (I'm not sure you can
use hdparm on their "Green" drives). That could be seen as "incompatibility"
but it's really WD's choice to support only Windows. That's a good reason
to avoid WD. If you end up with a set of hardware that manifests this bug
it's not really a big deal, and can be worked around easily with hdparm.

Frankly, you can avoid all these kinds of things by using last
generation hardware for Linux. Older stuff is well supported and Linux
doesn't need the latest hardware to run rings around Windows systems.

[Don Juan]

>
>
> Seems pretty much a non-issue. It's not because of Linux (you're
>
> just reading about it on a kernel forum :-), but because of a buggy interaction
>
> between certain laptop IDE drives and power management. It seems to depend
>
> on the installation platform as well as the particular drive since manifestation
>
> is dependent on IO patterns.
>

Uhm I agree, it isn't a kernel's bug, or a linux issue, but for what I've understood seems a non-standard use of ata commands (probably specific vendor extensions).
In particular with newer hdds, to obtain best performance in power saving and to make them more shock resistent head's device should be parked when disk is unused for a time.
In practice, the result is that if s.o. is booted by the hdd it (in transparent manner , to user) continuosly accesses to hdd.
The Device doesn't understand correctly commands relatively poer save management, has false beliefs on the world, and it cycles to parks unparks the heads (with deleterious effects on hdd's life).

>
>
> The issue here with WD is that WD chooses to require their own
>
> special program to reset APM parameters on a few models (I'm not sure you can
>
> use hdparm on their "Green" drives). That could be seen as "incompatibility"
>
> but it's really WD's choice to support only Windows. That's a good reason
>
> to avoid WD.

of course, after this experience, I will avoid them.
The Support center has told me simply:" here there are utilities for your model, with which you can diagnose the hdd health (note: I've just told them smart count load/unload, make diagnose wasn't my problem), there aren't fix for firmware, for your model there aren't software to tune this problem, but there is for others models, and has gave me the link, but also told me that this was not supported by WD ( looking on web, other users have reported that seems don't work)

> If you end up with a set of hardware that manifests this bug
>
> it's not really a big deal, and can be worked around easily with hdparm.

ok, is correct , in my experience I've used this utility and its .conf file, stopping the autopark function.
But honestly, I'm not sure of the consequence on policy's device power save managing (certainly, when the machine goes in stand-by, also hdd goes, but I seems that hdd can make a smart use of power, for example when it's not fully used)

>
> Frankly, you can avoid all these kinds of things by using last
>
> generation hardware for Linux. Older stuff is well supported and Linux
>
> doesn't need the latest hardware to run rings around Windows systems.

I'm not sure to understand your suggest (this only because English isn't my language, excuse me): Yuo suggest me, to avoid these problems, to prefer old hardware (probably without the cited smart parking feature )