On Thu, Nov 5, 2015 at 12:05 PM, D. Hugh Redelmeier <[email protected]> wrote:
> | From: o1bigtenor <[email protected]>
>
> | A tip - - - some of the drives marketed as applicable to NAS raid arrays 
> really
> | aren't applicable. You need to be purchasing drives that have ERC or error
> | recovery control. I was slapped upside the head because I had drives that 
> didn't
> | have that but when I bought the drives (early 2012) NOBODY was talking about
> | that.
>
> I've heard that some RAID systems do know how to deal with such
> drives.  I haven't researched which ones.
>
> Lots of us were talking about this (whining, actually).  I usually
> called it TLER, Western Digital's term for it.  ERC is Seagate's name.
> CCTL was used by Samsung and Hitachi.
>
> Earlier it was possible to tell a drive to limit error recovery time.
> Then the drive manufacturers locked this feature out on their cheap
> drives.  Grrr.

This sounds like a problem looking for a solution -- - - a software
hack to fix a
hardware problem - - - I know 0 about programming or I would be looking into
it already!
>
> Disks with TLER / ERC / CCTL & LCC [Table of drives] (2011 March)
> <http://hardforum.com/showthread.php?t=1590200>
>
> I posted to this list:
>         From: D. Hugh Redelmeier <[email protected]>
>         To: [email protected]
>         Date: Thu, 28 Jun 2012 21:58:59 -0400 (EDT)
>         Subject: Re: [TLUG]: 3TB Harddisk sale
>
>         | From: Anthony de Boer <[email protected]>
>
>         | I expect I'll be trying the WD RE4 Lennart mentions next.
>
>         My understanding is that RE and non-RE are the same EXCEPT for "TLER"
>         (a trivial firmware difference).
>
>         Without TLER, RAID won't work.  A drive will spend so much time
>         recovering from a simple local error that the controller will declare
>         the whole drive offline.  That is a big failure.  It generally
>         requires the array to be rebuilt, possibly taking longer than the
>         actual MTBF!
>
>         That's how they do "market segmentation".  Market segmentation is a
>         vendor's dream: sell essentially the same product at two different
>         price points.
>
>         If the drive manufacturing industry were not an oligopoly, this price
>         differentiation would disappear.  In fact, I think Samsung's normal
>         drives were capable of TLER; that's been fixed by Seagate taking them
>         over.
> ---

Very interesting - - - time for some hacker ingenuity!!!

Anyone that knows this stuff care to hand out some pointers?

Dee
---
Talk Mailing List
[email protected]
http://gtalug.org/mailman/listinfo/talk

Reply via email to