[mythtv-users] 2 TB Hard Drive Recommendations
beww at beww.org
Mon Dec 6 14:34:02 UTC 2010
On Monday, December 06, 2010 07:29:32 am jrlowry376 at comcast.net wrote:
> > Hi
> > On 6 December 2010 21:41, Andre <mythtv-list at dinkum.org.uk> wrote:
> >> I'm sure it's possible to have one track that's low density and
> >> readable without all the clever trickery usually used, then a
> >> minimal firmware bootstrap loader that knows only this track
> >> could load up the full firmware. Sounds like a common practice,
> >> should be ok provided the bootstrap loader knows how to re-write
> >> the firmware track!
> > I'm not sure if you've ever designed embedded firmware before (btw,
> > that's what I do for a living), but the amount of smarts required
> > to start spinning the disk, move the head to the right spot, read
> > the data, load the firmware in RAM and start is quite massive to
> > start with.
> > As you said, you would need a firmware for just that (probably on
> > flash). Constraints on flash make them all of a minimum size to
> > start with.
> > The drive needs a writing location to record the bad sectors and
> > other bad spot detected on the disk too..
> In the '80s I worked on Data General Argus disk drives, there where
> three versions with capacities of 354, 592 and 862 Megabytes. I
> can't remember if they where 6809 or Data General micro Nova based
> I do remember that the firmware was stored on the drive, and was
> easily upgraded by running diagnostics on the attached super mini
> computer. There where 16 spare sectors per cylinder, remapping was
> also a diagnostic activity. If you ran out of spare sectors in a
> cylinder you replaced the HDA.
> We've come a long way from those drives (in a lot of ways) these
> drives where 19" wide, 10.5" tall and 20+" deep and required two
> engineers to replace a complete drive. Three of the 354s where sold
> as package in a meter high rack as a Giga-stack.
> I'll agree that it may take a lot of code to run internal
> diagnostics, spin a drive and read in firmware, but it's a subset of
> the code required to run the drive in normal operation.
> I'm not a bean counter, but I'm sure there was a lot of thought put
> into the decision to store the firmware on the drive.
> With all of this talk about drive reliability it's useful to read the
> Google drive reliability report.
I read it way back when, I seem to recall they did not mention any
specific manufacturers, which is understandable, but it makes it harder
to make a buying decision.
The biggest surprise I recall was that high temperatures were not as big
a contributor to failures as most people thought.
More information about the mythtv-users