[mythtv-users] Fwd: Further Notice of Seagate Hard Drive Class Action and Proposed Settlement

[John Drescher]

On Fri, Mar 12, 2010 at 2:58 PM, Brian Wood <beww at beww.org> wrote:
> On Friday 12 March 2010 12:37:12 pm Ben Kamen wrote:
>> On 3/12/2010 1:16 PM, Brian Wood wrote:
>> > What about the extra power consumed by the extra bit in ECC RAM, and the
>> > power used by the buffer chips?
>> >
>> > Certainly there should be some sort of disclaimer of these factors for
>> > all ECC RAM.
>> >
>> > Plus, I was induced to buy ECC RAM because the "error correcting" feature
>> > appealed to me as a bad typist. I have been deceived by marketing, since
>> > I still see errors in my documents that are stored in such RAM.
>> >
>> > Legislation is Urgently Needed!
>>
>> And how many times have you actually seen an OS or BIOS that can
>>  successfully cope with DRAM errors when one has ECC/Parity RAM installed?
>>
>> I used to fix IBM PC's back in the 80's when they all came with Parity RAM.
>
> Quite true, all RAM back then had 9 bits. They eliminated the "extra" bit
> either because they realized it was useless, or because they figured out they
> were missing a trick to make more money.
>
>>
>> I never once saw a computer that could cope with RAM going bad and parity
>>  protecting it.
>
> I guess theoretically you wouldn't notice it until it got so bad that ECC
> couldn't cope with it.
>

Server motherboards log this. I had an opteron mobo that recorded a
correctable ECC error about 2 times a week. It was fully stable as
long as I did not run it with dual 285s.

John