[wellylug] USB 3.0 via PCI-e, & SATA HDD enclosures/docks

[Cliff Pratt]

On 18/07/11 19:58, Daniel Reurich wrote:
> On 16/07/11 16:33, Cliff Pratt wrote:
>> On 15/07/11 23:41, Ewen McNeill wrote:
>>> On 2011-07-15 20:48 , Cliff Pratt wrote:
>>>> I've removed many (dead) still spinning disks and I've noticed the
>>>> weird gyro effect of tilting a fast spinning disk....
>>>
>>> My general approach for hotswap disk trays is to pull the disk out
>>> enough that it's no longer connected, and then leave it a bit (eg, 30
>>> seconds) for the platters to spin down before doing anything else
>>> with it. I guess (for David's situation) that deliberately telling
>>> the disk to spin down first might save some time there. But such
>>> software commands feel very 1980s "park the heads" like to me...
>>>
>> I agree, in the sense that it was useful at one time, but outlived its
>> usefulness by several years.
>>
> I can't believe this approach is being entertained. It sounds like a
> recipe for disaster and may result in early failure of the drives.
>
> Hotplug is a feature that allows for the adding and removal of devices
> at an electrical level without causing damage to the componentry and
> logic circuits and sends the OS events on the attaching/removal of
> devices. It does nothing for protecting the storage medium from failure
> due to badly timed removal or protecting the information being
> stored/retrieved, especially given the amount of cache that spinning
> rust drives have these days.
>
> I don't think it's wise to assume that a drive will just cope with
> having it's comms and power pulled without preparing it for removal
> first, including giving it opportunity to spin down platters and park
> heads. Given the ever increasing data densities and corresponding
> reduction in acceptable operating tolerances, one can only assume that
> physically moving a drive in whilst the platters are spinning and the
> heads still hovering over the data area is only going to add to wear and
> tear if not cause some more serious damage.
>
> Saying that parking the heads is archaic does not wash IMHO. It just so
> happens that in normal circumstances now the OS and drive firmware
> combinations are smart enough to ensure that happens at appropriate
> times. I don't consider physical removal of a drive without giving the
> OS and drive forewarning is considered as normal and expected behavior.
>
> It might be somewhat safer for the latest SSD drives which have
> capacitors that may hold enough power to will the cache to non-volatile
> storage, but even then some preparation would still be a wise approach
> to ensure data integrity.
>
> My $0.02 cents worth :-)
>
Yeah good points, but a) I thought that drives/caches retain enough 
power to write the last of the data in the cache and b) why don't 
hotplug devices come with dire warnings and why aren't there 'spindown' 
programs (GPL or otherwise!)

I was also under the impression that even older IDE drives didn't 
actually park the heads. I thought that when the juice is removed the 
spin of the disks and the lack of power automatically raised the heads 
to a safe position?

Cheers,

Cliff