Return-Path: <foner@media.mit.edu>
Received: from pacific-carrier-annex.mit.edu by po11.mit.edu (8.9.2/4.7) id BAA26514; Mon, 20 May 2002 01:57:14 -0400 (EDT)
Received: from out-of-band.media.mit.edu (out-of-band.media.mit.edu [18.85.16.22])
	by pacific-carrier-annex.mit.edu (8.9.2/8.9.2) with ESMTP id BAA06433;
	Mon, 20 May 2002 01:56:32 -0400 (EDT)
Received: (from foner@localhost)
	by out-of-band.media.mit.edu (8.9.3/8.9.3/MEDIA) id BAA07918;
	Mon, 20 May 2002 01:56:29 -0400 (EDT)
Date: Mon, 20 May 2002 01:56:29 -0400 (EDT)
Message-Id: <200205200556.BAA07918@out-of-band.media.mit.edu>
From: Lenny Foner <foner@media.mit.edu>
To: nathanw@MIT.EDU
CC: marc@MIT.EDU, tytso@MIT.EDU, jmorzins@MIT.EDU, sipb@MIT.EDU
In-reply-to: <mtu4rh332wn.fsf@department-of-alchemy.mit.edu> (nathanw@MIT.EDU)
Subject: Which hard disk vendor do you trust?
CC: foner@media.mit.edu

[Err, uhm, if this is still considered SIPB-appropriate, fine,
 but I'd rather not spam the list if others would rather that
 the discussion go elsewhere...]

    Date: 19 May 2002 23:37:44 -0400
    From: nathanw@MIT.EDU (Nathan J. Williams)

    Lenny Foner <foner@media.mit.edu> writes:

    > Also obviously, it cannot be the case that TiVos are simply
    > dying in 3-6 months, or TiVo wouldn't be in business any more.

    I suspect that you are unlucky. TiVos do have disk failues, even with
    the disks shipped from the factory; digging around on any of the
    message boards will show that.  TiVoS are an application where a
    nonzero, even high failure rate pretty much *is* acceptable - covered
    with the insurance policy of a warranty - and the cost savings from
    that are necessary to produce a salable product.

Yes, but you're missing the point, as have many people in this
discussion.  It's clear that -some- failures happen, but I'm seeing
them in 3-6 months, consistently.  The TiVo parts & labor warranty is
only 3 months anyway.[1]  Customers would be screaming bloody murder
if IDE drive failures were common with anything like that frequency.
Thus, I suspect that Maxtor is not the right vendor for this
application.  Therefore, -who is-?

That's been my question from the very start.

I frankly -do not care- about discussions comparing the merits of IDE
vs SCSI reliability, and I'm sorry that I said anything to mislead
anyone into thinking that I did.  Whether or not it's actually true
(and I'm still waiting for data, not just assertions), it -cannot help
me- if I'm using something like a TiVo, because a TiVo -cannot use
SCSI disks-.  (It also can't help my machine in colo, since neither
Sun Netras nor Sun Fire V100's use SCSI disks either. IDE only.  Hmm.)

Therefore, of -only- IDE disks in 80GB or higher capacities,
which vendor(s) seem to have the best reliability track record?

Sheesh.  It's -amazing- how hard it's been to keep people on track
in this discussion.  Everyone seems to want to wander away into
SCSI la-la land instead of addressing the original question.  I'm
sorry to sound rather cross, but I've gotten, oh, about 10 or so
similar responses (most of which didn't go to SIPB; many of which
didn't come from there) which all sound the same and which all miss
the entire point of the original question.  I blame myself.  Clearly,
I wasn't nearly forceful enough in the first place that opinions on
-IDE- drives were the only ones that would be helpful to me.  My guess
is that many of the people who are responding don't actually have any
sort of well-grounded opinion on IDE drive vendors (much less actual
data collected via some sort of controlled process), but feel compelled
to say -something-, so they talk about SCSI.  That's just not helpful.

[1] TiVo's ship with Quantums, but typically only 15GB.  If I can
believe what at least one other person has said, that doesn't say
anything, because they claimed that small disks were inherently more
reliable.
