[wellylug] USB 3.0 via PCI-e, & SATA HDD enclosures/docks
Daniel Reurich
daniel at centurion.net.nz
Mon Jul 18 19:58:09 NZST 2011
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 :-)
--
Daniel Reurich
Centurion Computer Technology (2005) Ltd
Ph: 021 797 722
-------------- next part --------------
A non-text attachment was scrubbed...
Name: daniel.vcf
Type: text/x-vcard
Size: 303 bytes
Desc: not available
URL: <http://lists.wellylug.org.nz/pipermail/wellylug/attachments/20110718/04270d76/attachment.vcf>
More information about the wellylug
mailing list