|
Home > Archive > Data Storage > May 2005 > ATA Reliability: Seagate, WD, Maxtor
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
ATA Reliability: Seagate, WD, Maxtor
|
|
|
| Not to start any heated disputes here, but I'd like to get general
feedback re reliability of the three major brands. I run a few Raid
arrays...Actually Raid 0, so reliability is important.
WD was the first to feature larger buffers, and many started using
them as a result of the p.r. surge. I'm not sure if their quality has
dropped, but I started having reliability problems with them about
a year ago. That was compounded by my experience with WD
tech support. Difficult to reach. Difficult to deal with. Your
experience may vary, of course, but if there *is* a problem it's
nice to have a smooth path to fixing it. So I started using Maxtor.
With regard to support, Maxtor has been much easier to deal with.
Their techs usually take time to track down problems, and they're
not hesitant to escalate to level 2 support when warranted.
But...I've had a few Maxtor failures recently, so... <g>
I've bought some Seagates in the hope that they're back near
the top (Seagate was *the* name once upon a time). I don't
have any experience with their support yet. I have no idea
how they stack up to WD and Maxtor yet. So I'm hesitant to
invest in a lot of drives until they're proven out.
Any opinions appreciated, especially with regard to Seagate.
_R
| |
| Charles Morrall 2005-04-06, 5:46 pm |
|
"_R" <_R@nomail.org> skrev i meddelandet
news:rf07515j78tv8gnhbrsprulqho8p7hmdib@
4ax.com...
> Not to start any heated disputes here, but I'd like to get general
> feedback re reliability of the three major brands. I run a few Raid
> arrays...Actually Raid 0, so reliability is important.
You value reliability, but you run Raid0? I understand you'd want reliable
drives because of this, but I would never run raid0 and hope the drive
doesn't break. If reliability is truly important, run raid>0.
/charles
| |
| Faeandar 2005-04-06, 8:45 pm |
| On Wed, 06 Apr 2005 02:51:06 -0400, _R <_R@nomail.org> wrote:
>Not to start any heated disputes here, but I'd like to get general
>feedback re reliability of the three major brands. I run a few Raid
>arrays...Actually Raid 0, so reliability is important.
>
>WD was the first to feature larger buffers, and many started using
>them as a result of the p.r. surge. I'm not sure if their quality has
>dropped, but I started having reliability problems with them about
>a year ago. That was compounded by my experience with WD
>tech support. Difficult to reach. Difficult to deal with. Your
>experience may vary, of course, but if there *is* a problem it's
>nice to have a smooth path to fixing it. So I started using Maxtor.
>
>With regard to support, Maxtor has been much easier to deal with.
>Their techs usually take time to track down problems, and they're
>not hesitant to escalate to level 2 support when warranted.
>But...I've had a few Maxtor failures recently, so... <g>
>
>I've bought some Seagates in the hope that they're back near
>the top (Seagate was *the* name once upon a time). I don't
>have any experience with their support yet. I have no idea
>how they stack up to WD and Maxtor yet. So I'm hesitant to
>invest in a lot of drives until they're proven out.
>
>Any opinions appreciated, especially with regard to Seagate.
>
>_R
>
As Charles pointed out, raid 0 is "zero raid". So if you've already
started and you're really using raid 0 you're in trouble.
As to ATA reliability, they aren't. That's why most reputable vendors
have maximum raid group sizes or use special parity when usng ATA
drives. Not to say they can't work well for you, just that burning
time and effort on ATA reliability is fruitless IMO. Just buy some
and make sure you've built your raid properly.
~F
| |
|
| In article <nts85155k5upgflrci7e6nfv5kmk3s5mma@4ax.com>,
Faeandar <mr_castalot@yahoo.com> wrote:
> As to ATA reliability, they aren't. That's why most reputable vendors
Actually, they *are* fairly reliable.
I'm not sure anyone vendor is better than the others. Also, you left out Hitachi.
> have maximum raid group sizes or use special parity when usng ATA
> drives. Not to say they can't work well for you, just that burning
Can you name any vendors that has this "maximum group size" and what is it?
> time and effort on ATA reliability is fruitless IMO. Just buy some
It's not in mine. But I would choose SATA over ATA because the SATA interface is better (e.g, no master/slave, etc).
| |
|
| On Wed, 06 Apr 2005 14:57:20 GMT, "Charles Morrall"
<charles.morrall@telia.com> wrote:
>
>"_R" <_R@nomail.org> skrev i meddelandet
> news:rf07515j78tv8gnhbrsprulqho8p7hmdib@
4ax.com...
>
>You value reliability, but you run Raid0? I understand you'd want reliable
>drives because of this, but I would never run raid0 and hope the drive
>doesn't break. If reliability is truly important, run raid>0.
>
>/charles
In a perfect world I could just keep throwing redundant drives at it.
The configuration is necessary given the system constraints.
There are two RAID 0 pairs per box; one for multitrack audio, the
other for handling video. Raid0 was necessary for throughput more
than for volume. They are on 3Ware controllers which I've found
reliable. And unlike the low-end Promise controllers, 3ware is
true raid (separate master channels per drive).
Systems are mirrored, but not in real time. (Slow Raid if you will)
If a bug manages to get through to someone's normal Raid array,
it's toast no matter what the redundancy. With scheduled backups
there is a chance that I could crash and lose data, but also the
chance of a virus taking down the mirror is diminished. It's a
calculated risk.
Anyway, Maybe I should not have emphasized Raid0 as it does
seem to touch a button with some. The rest of the drives in
each system run non-Raid. And when sustained bandwidth
of drives increases, we will drop RAID0.
I'm more interested in reliability of individual drive manufacturers
and models than in Raid0. And there are differences. Even
firmware quality (WD crashes) can enter into it. And as I
pointed out, tech support after the fact is important.
No comments on Seagate, eh?
_R
| |
| Faeandar 2005-04-07, 5:59 pm |
| On Thu, 07 Apr 2005 03:42:44 GMT, flux <support@fluxsoft.com> wrote:
>In article <nts85155k5upgflrci7e6nfv5kmk3s5mma@4ax.com>,
> Faeandar <mr_castalot@yahoo.com> wrote:
>
>
>Actually, they *are* fairly reliable.
>
>I'm not sure anyone vendor is better than the others. Also, you left out Hitachi.
I'm guessing this is for the OP and not me since I'm not sure how I
left out Hitachi.
Fairly reliable is not the same thing. Do you want an airplane that's
fairly reliable? Point is that reliability of ATA is nowhere near
that of Scsi or FC. Once you get down into the desktop class drives
my view on it is rely on raid. Of course that's my view on the high
end too so....
But I also don't care about who makes the high end drives either. If
I have alot of a specific vendor fail then I may start asking my
vendor about it, but otherwise it's just swap and go.
>
>
>Can you name any vendors that has this "maximum group size" and what is it?
I should have said a recommended maximum raid group size. Nothing
keeps me from shooting myself if I so choose.
NetApp
HDS
HP
IBM
Although in some of the above cases it's SATA and not ATA. But SATA
is slightly more reliable than ATA which is some don't even use ATA at
all.
~F
| |
|
| In article <sbha51542cjstjd33t4tnb5n55l8okmv9n@4ax.com>,
Faeandar <mr_castalot@yahoo.com> wrote:
> Fairly reliable is not the same thing.
It's not the same thing as what?
> Do you want an airplane that's
> fairly reliable?
Yes.
> Point is that reliability of ATA is nowhere near
> that of Scsi or FC. Once you get down into the desktop class drives
This statement is just fiction. Drives today are roughly the same in reliability, marketing claims notwithstanding.
> I should have said a recommended maximum raid group size. Nothing
Which is?
| |
| Faeandar 2005-04-08, 2:45 am |
| On Fri, 08 Apr 2005 04:11:11 GMT, flux <support@fluxsoft.com> wrote:
>In article <sbha51542cjstjd33t4tnb5n55l8okmv9n@4ax.com>,
> Faeandar <mr_castalot@yahoo.com> wrote:
>
>
>It's not the same thing as what?
Not the same as reliable. Fairly reliable != reliable.
>
>
>Yes.
Good luck with that. I want one that's very reliable.
>
>
>This statement is just fiction. Drives today are roughly the same in reliability, marketing claims notwithstanding.
Too many vendors discuss ATA and SATA as tier2 storage for this to be
fiction. Pick a vendor, talk to them about a SATA or ATA array and
see what they say. Also ask them about their failure rates for each
type of drive. 4 vendors I've talked to all say the same thing,
(S)ATA are tier2. Both for reliability and performance.
>
>
>Which is?
Depends on the vendor. NetApp is 8, HDS is 10 (depending on what
category you get), and I believe IBM is 10 also but can't recall
exactly. A noteworthy point too is that of these only NetApp uses
ATA, the others use SATA.
~F
| |
|
| On Fri, 08 Apr 2005 04:11:11 GMT, flux wrote:
> Faeandar wrote:
>
>
>This statement is just fiction. Drives today are roughly the same in reliability, marketing claims notwithstanding.
No - I think Faeandar has a point. It all comes down to cost.
FC and SCSI drives are intended for Enterprise-class use; they are
manufactured, tested and priced accordingly.
Generally speaking, ATA and SATA drives are intended for desktop use
and although they may be designed in a similar (or even the same) way,
they are not subjected to the same testing regime or manufacturing
tolerances as Enterprise-class products.
Some vendors offer Enterprise type SATA drives - generally with faster
RPM speeds and higher prices. I would expect these drives to have had
more thorough testing than a standard desktop item.
The vendors that are using ATA/SATA in products aimed at Enterprise
use, like NetApp, have taken additional steps to ensure reliability.
This could be stricter use of SMART data, probes on the data path or
additional parity, specifically intended to improve the reliability of
ATA/SATA subsystems.
HVB.
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-08, 5:46 pm |
| In article <84ec51d5ce0bbv6mnjvsfcq0qtppu7diej@4ax.com>,
HVB <devnull@127.0.0.1> wrote:
>...
>Generally speaking, ATA and SATA drives are intended for desktop use
>and although they may be designed in a similar (or even the same) way,
>they are not subjected to the same testing regime or manufacturing
>tolerances as Enterprise-class products.
>...
No, wrong. Desktop-class drives are DESIGNED radically different from
Enterprise-class drives. This has been true for about a decade. The
last time SCSI and IDE drives were the same underlying drive with just
different interface boards was a long time ago. For the gory detail
of the massive differences between desktop-class and enterprise-class
drives, start by reading: "More than an Interface - SCSI versus ATA",
by Erik Riedel, Jim Dykes, and Dave Anderson, available at a web
search near you.
There is a huge difference in cost, performance characteristics
(tradeoff between capacity and speed), and reliability between
desktop-class and enterprise-class drives. In a nutshell, one could
say: desktop class drives are very cheap, have very high capacity, but
they are slow, and unreliable (both in overall livetime, and also in
their resilience to problems, like they don't like to deal with
vibration). If you look how they are engineered differently (single
combined servo/datapath processor, meaning unable to servo the head
while writing, slower spindles, weaker actuators, fewer air filters,
larger platters, lightweight but weaker frames), this all makes sense.
Little of the difference between desktop-class and enterprise-class
drives is about testing. The story is not that you start with
fundamentally the same drive, and the ones that pass the test get a
SCSI board bolted on and are sold for $$$, while the ones that fail
the test get an ATA board bolted to it and are sold for $.
Internally, the design is radically different.
Now, what is true: While all FC/SCSI drives are enterprise-class
drives, not all ATA/SATA drives are desktop-class drives. About 2
years ago, some manufacturers (names withheld to protect the guilty)
started a trend of selling purported enterprise-class drives with ATA
interfaces. Today, there are quite a few supposedly enterprise-class
drives being sold with SATA interfaces. Unfortunately, I haved talked
to experts in the field (names withheld), which have performed a
teardown analysis on some (but not all!) of these ATA/SATA
enterprise-class drives, and they find that they are built like
desktop-class drives; this was particularly true of the early models.
The following rule of thumb seems to hold in many cases: If something
is nearly as cheap (in $ per byte) as a desktop-class drive, it is
unlikely to be a reliable enterprise-class drive. You do get what you
pay for. If you want a free lunch, look elsewhere.
Now, naturally you can take low-reliability drives, and using
RAID-style techniques built high-reliability disk systems out of them.
The industry has been doing this since the late 70s or early 80s (even
though the buzzword RAID was only coined in 1989). Given the
extremely large capacity of modern drives, and their dropping
reliability (the actual rate of loss of bytes is increasing, because
the capacity is increasing much faster than the reliability), more and
more exotic RAID techniques are required these days (a dumb RAID 5
with a large group, without scrubbing and/or failure prediction is
unlikely to cut it any longer). Whether the building of RAID arrays
by amateurs using off-the-shelf commodity components and inexpensive
disks is a good value, everyone has to determine for themselves,
making a tradeoff between cost of goods, cost of effort for building a
disk system, and value of the data = cost of data loss.
Disclaimer: My employer (which shall remain nameless) does not
manufacture disk drives, but uses many of them, and I do storage
systems for a living. But at home I use a mix of enterprise-class
drives in a RAID configuration (for stuff I care about, like baby
pictures) and cheap drives bought at low-end computer stores (for
stuff downloaded from the net and for backups).
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
|
| On Fri, 08 Apr 2005 16:51:44 -0000,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>In article <84ec51d5ce0bbv6mnjvsfcq0qtppu7diej@4ax.com>,
>HVB <devnull@127.0.0.1> wrote:
>
>No, wrong. Desktop-class drives are DESIGNED radically different from
>Enterprise-class drives.
I know that. In my defense... I did say *may be designed* not are
designed.
HVB
| |
| Paul Rubin 2005-04-08, 5:46 pm |
| _R <_R@nomail.org> writes:
> Not to start any heated disputes here, but I'd like to get general
> feedback re reliability of the three major brands. ...
> Any opinions appreciated, especially with regard to Seagate.
Yes, there are differences between brands. But those differences
change from week to week. There are often specific models that are
turkeys while other similar models from the same manufacturer are
good. Yes you can buy enterprise drives and get more reliability than
desktop drives, but that's not simply a matter of spending more money
and getting more reliability with no other cost except to your wallet.
The enterprise drives have lower capacity per spindle, need more power
and therefore more cooling, and make more noise. If you unplug a
desktop drive and replace it with an enterprise drive without adding
more cooling, the result can often be LESS reliable than what you had
before.
As for desktop drive reliability, AFAIK the dynamic is something like
this. A manufacturer designs a new type of drive conservatively, but
being a new design it has some bugs so it's not so reliable. Then the
bugs get fixed, so it's more reliable. Then the design enters the
usual cycle, where there's relentless pressure over time to increase
capacity and lower costs. The technology in the design gets pushed
closer and closer to its limits and reliability suffers. Eventually
the drives from that product line become so unreliable that the return
rate becomes unacceptable and also the manufacturer takes a beating in
the marketplace. At that point they have to clean up their act, which
can mean coming up with another design that's technologically newer
than the old one, so it can more conservative again in terms of how
far it presses the technology's capability. The famous IBM/Hitachi
Deskstar (a/k/a Deathstar) debacle went something like this, I think.
Of course none of this is reflected in the marketing crap that the
front office issues. The "New! Hyper-whizbang XYZ" drive can just be
another iteration of the "Old! Bogo-fizzle ABC" drive with a few more
GB and a few cents of cost reduction and less reliability. While the
big engineering change is actually between the "Click-whir 23A-24518"
and the "Click-whir 23A-24519".
| |
| Curious George 2005-04-09, 2:45 am |
| On Fri, 08 Apr 2005 18:24:04 +0100, HVB <devnull@127.0.0.1> wrote:
>On Fri, 08 Apr 2005 16:51:44 -0000,
>_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>
>
>I know that. In my defense... I did say *may be designed* not are
>designed.
>
>HVB
I think it's kinda funny how in other groups this would normally
quickly turn into a flame war spurred by some user(s) who say, "my
desktop ATA drive works fine. What are you talking about? All drives
are the same except SCSI is a rip-off!" with maybe 1 person
countering on the ES side. While in a group of basically storage
professionals like this the ES/PS distinction is barely contentious;
only its details are to a limited extent.
For the OP FWIW I have noticed greater infant mortality, higher
incidence of motors which turn noisy, and higher likelihood of
defective drives which allow wild writes as well as just general
inconsistency across the ES & PS product line with Maxtor drives over
the last decade as opposed to most other brands. The problem,
however, might be more related to my supplier & delivery route than
actually Maxtor's fault (I don't know). Careful handling, integration
& monitoring are probably more important for success than brand name.
Everyone occasionally makes a bum model. Your or my bad luck may not
be indicative of the overall brand.
There is an interesting somewhat anecdotal project over at
storagereview.com- their reliability database. I try to mention it
when I can because the more ppl that contribute the better chance we
have of it yielding accurate results. (Please not that I am not
affiliated in any way with storagereview.com or its sponsors)
| |
| Paul Rubin 2005-04-09, 2:45 am |
| Curious George <cg@email.net> writes:
> For the OP FWIW I have noticed greater infant mortality, higher
> incidence of motors which turn noisy, and higher likelihood of
> defective drives which allow wild writes as well as just general
> inconsistency across the ES & PS product line with Maxtor drives over
> the last decade as opposed to most other brands.
I had a weird and somewhat scary experience with a Maxtor 60gb drive.
I bought it about 4 years ago, mounted it in a PC, installed an OS on
it and did a couple things on it. That all worked fine. The machine
was then powered off and ended up being taken out of service. Three
years later I decided to use the drive for something else. The drive
had no more than a dozen or so power-on hours and had just been
sitting in an unused PC for 3 years. It spun up normally but I could
no longer read data from it and the connected host didn't recognize it.
This is a little bit disturbing since it makes using HD's as archive
media sound unreliable. I asked someone about it and as soon as I
said "60 GB drive" he immediately recognized the problem. He said
that there was some contamination problem in an ALPS factory that made
the heads for basically all 60GB drives of that era. And so the same
thing likey would have happened no matter what brand of drive I'd
bought, but wouldn't have happened with a 40gb drive or an 80gb drive.
Gack. I still don't know my long term archive strategy. I guess LTO
still looks best.
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-09, 2:45 am |
| In article <ar5e51lhp87pfu4fj2ciqbu0865q8uqcd9@4ax.com>,
Curious George <cg@email.net> wrote:
>On Fri, 08 Apr 2005 18:24:04 +0100, HVB <devnull@127.0.0.1> wrote:
>
>
>I think it's kinda funny how in other groups this would normally
>quickly turn into a flame war ...
In retrospect, my very terse "no, wrong" sentence was less than
polite. I apologize for being too curt.
>There is an interesting somewhat anecdotal project over at
>storagereview.com- their reliability database. I try to mention it
>when I can because the more ppl that contribute the better chance we
>have of it yielding accurate results. (Please not that I am not
>affiliated in any way with storagereview.com or its sponsors)
Most certainly all systems integrators that use large quantities of
disks (ranging from Dell at the consumer end to HDS at the enterprise
end) carefully evaluate the data on reliability of drives, and keep
tabs on failures rates in the field. Certainly the data is correlated
not just with manufacturer info, but also model, age, serial number
range, and (where known) environmental conditions. And I'm also sure
that all these companies keep that data closely guarded; not only is
it a useful trade secret, but if leaked out it could also be lawsuit
material.
Think about it this way. If it weren't for Consumer Reports, how
would you ever find out whether Ford's or Chevy's are more reliable?
You could call both companies headquarters and ask for statistics -
good luck! You could ask your beer-drinking buddies, and all you'd
get were religiously held opinions, and a few anecdotes. Getting real
information on such an important question is just very hard.
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-09, 2:45 am |
| In article <7xsm20wtrs.fsf@ruckus.brouhaha.com>,
Paul Rubin <http://phr.cx@NOSPAM.invalid> wrote:
>Curious George <cg@email.net> writes:
>I had a weird and somewhat scary experience with a Maxtor 60gb drive.
>I bought it about 4 years ago, mounted it in a PC, installed an OS on
>it and did a couple things on it. That all worked fine. The machine
>was then powered off and ended up being taken out of service. Three
>years later I decided to use the drive for something else. The drive
>had no more than a dozen or so power-on hours and had just been
>sitting in an unused PC for 3 years. It spun up normally but I could
>no longer read data from it and the connected host didn't recognize it.
>
>This is a little bit disturbing since it makes using HD's as archive
>media sound unreliable. I asked someone about it and as soon as I
>said "60 GB drive" he immediately recognized the problem. He said
>that there was some contamination problem in an ALPS factory that made
>the heads for basically all 60GB drives of that era. And so the same
>thing likey would have happened no matter what brand of drive I'd
>bought, but wouldn't have happened with a 40gb drive or an 80gb drive.
It seems unlikely that all manufacturers of that era (Seagate,
IBM/Hitachi, Maxtor=Quantum, WD, plus the Korean ones that make ATA
only) would all use ALPS heads. ALPS has a high market share, but I
can't really imagine it being near 100%.
>Gack. I still don't know my long term archive strategy. I guess LTO
>still looks best.
There are persistent rumors around that the lubrication in the
bearings (both spindle and actuator) will do something nasty like turn
into tar, if you store a drive for a long time without running it. So
relying on non-spinning drives for long-term archive seems dangerous.
At least not without asking the manufacturer of the drive whether they
support this mode.
For my stuff at home, I have two disk copies of everything that's in
use, plus one copy of all the archival stuff on a single disk. So far
increasing disk capacity has made using a single disk as an archive
viable (it is currently a 200GB ATA disk). Plus one extra copy of
everything (most importantly the archive) on writeable CDs.
Obviously, things that can be easily recreated (like tracks ripped
from audio CDs) are exempt from the archive and backup. But it is
quite amazing: For a home computer, including digital pictures (but
not digital video), the increase in disk capacity over the last 10
years has meant that I never have to delete anything.
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
|
| In article <1112979104.459192@smirk>,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
> No, wrong. Desktop-class drives are DESIGNED radically different from
> Enterprise-class drives. This has been true for about a decade. The
Practically speaking, this is not so.
> There is a huge difference in cost, performance characteristics
> (tradeoff between capacity and speed), and reliability between
> desktop-class and enterprise-class drives. In a nutshell, one could
In other words, marketing.
> say: desktop class drives are very cheap, have very high capacity, but
> they are slow, and unreliable (both in overall livetime, and also in
Cheap yes, high capacity yes.
Slow no, unreliable no.
> their resilience to problems, like they don't like to deal with
> vibration). If you look how they are engineered differently (single
In fact, they are very resilient to vibration.
> Internally, the design is radically different.
Where is the evidence?
> drives being sold with SATA interfaces. Unfortunately, I haved talked
> to experts in the field (names withheld), which have performed a
> teardown analysis on some (but not all!) of these ATA/SATA
Let's have the names.
> Now, naturally you can take low-reliability drives, and using
Are there any such drives any more?
> manufacture disk drives, but uses many of them, and I do storage
> systems for a living. But at home I use a mix of enterprise-class
I'm not sure I believe you.
| |
|
| In article <84ec51d5ce0bbv6mnjvsfcq0qtppu7diej@4ax.com>,
HVB <devnull@127.0.0.1> wrote:
> On Fri, 08 Apr 2005 04:11:11 GMT, flux wrote:
>
>
> No - I think Faeandar has a point. It all comes down to cost.
>
> FC and SCSI drives are intended for Enterprise-class use; they are
> manufactured, tested and priced accordingly.
But manufacturers sell few FC/SCSI drives compared to ATA drives, so it
makes sense for there to be cost difference.
> Generally speaking, ATA and SATA drives are intended for desktop use
> and although they may be designed in a similar (or even the same) way,
> they are not subjected to the same testing regime or manufacturing
> tolerances as Enterprise-class products.
>
> Some vendors offer Enterprise type SATA drives - generally with faster
> RPM speeds and higher prices. I would expect these drives to have had
> more thorough testing than a standard desktop item.
>
> The vendors that are using ATA/SATA in products aimed at Enterprise
> use, like NetApp, have taken additional steps to ensure reliability.
> This could be stricter use of SMART data, probes on the data path or
> additional parity, specifically intended to improve the reliability of
> ATA/SATA subsystems.
I don't know how much of it is true, but I do know that ATA drives are a
lot faster and a lot more reliable than some here think they are.
| |
| Curious George 2005-04-09, 2:45 am |
| On Sat, 09 Apr 2005 04:51:02 -0000,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>Most certainly all systems integrators that use large quantities of
>disks (ranging from Dell at the consumer end to HDS at the enterprise
>end) carefully evaluate the data on reliability of drives, and keep
>tabs on failures rates in the field.
And if you're a small fish, or any kind of consumer you're totally out
in the cold.
>Think about it this way. If it weren't for Consumer Reports, how
>would you ever find out whether Ford's or Chevy's are more reliable?
Wouldn't it be great if this type of nfo existed for all technology
products? And from a 3rd party not potential who knows who with who
knows what axe to grind?
>You could call both companies headquarters and ask for statistics -
>good luck!
Oh sure you'll get "reliability numbers"; mendacious drivel like
theoretical MTBF.
> You could ask your beer-drinking buddies, and all you'd
>get were religiously held opinions, and a few anecdotes.
Like most of the usenet & board posts on the subject
(I fully admit the limitations of my Maxtor comment)
> Getting real
>information on such an important question is just very hard.
Really Really Agree!
| |
| Bill Todd 2005-04-09, 7:45 am |
| flux wrote:
....
> I'm not sure I believe you.
Then you might benefit from actually looking at the material he cited
rather than pulling opinions directly out of your XXX.
- bill
| |
| Bill Todd 2005-04-09, 7:45 am |
| flux wrote:
....
> I don't know how much of it is true,
Rather obviously.
but I do know that ATA drives are a
> lot faster and a lot more reliable than some here think they are.
Since I don't recall anyone yet having actually quantified either value,
your confidence in that statement seems rather unfounded.
Fact: ATA drives max out at 7200 rpm, while SCSI/FC drives max out at
15,000 rpm. That gives SCSI drives a 2+:1 advantage in rotational
latency right off the bat.
(Yes, WD's Raptor spins at 10,000 rpm, which reduces the SCSI/FC
advantage to 1.5:1. But you referred to 'ATA', not 'SATA', above.)
Fact: ATA drives have average seek latencies in the 8 - 9+ ms. range
(there may be one in the 7s somewhere), while the fastest SCSI/FC drives
have average seek times less than half that (I haven't bothered to check
out the WD Raptor in this regard). Another 2+:1 speed advantage for
SCSI/FC.
Fact: Unlike SCSI/FC drives, ATA drives do not typically support
command queuing that allows seek optimization at the drive to further
reduce average access latencies (and even those ATA drives that do often
lack drivers that support this feature). Advantage: varies with the
queue depth, but typically quite significant in enterprise use (i.e.,
when parallel rather than serial access predominates) and always at
least noticeable as long as the queue depth exceeds 1.
As for reliability, I'll let you read the paper that you previously
neglected to: it's quite thorough in its assessment of the differences.
- bill
| |
|
| On Sat, 09 Apr 2005 02:39:08 GMT, Curious George <cg@email.net> wrote:
>On Fri, 08 Apr 2005 18:24:04 +0100, HVB <devnull@127.0.0.1> wrote:
>
>I think it's kinda funny how in other groups this would normally
>quickly turn into a flame war spurred by some user(s) who say, "my
>desktop ATA drive works fine. What are you talking about? All drives
>are the same except SCSI is a rip-off!"
BULLSHIT!!! <g>
> with maybe 1 person
>countering on the ES side. While in a group of basically storage
>professionals like this the ES/PS distinction is barely contentious;
>only its details are to a limited extent.
Are there other storage groups? I only know of this one.
I don't put myself in the 'SCSI-only, or server' school, but I have
terabytes of storage (I would be broke if it were SCSI). Much of it
is for archive purposes. Some of it is on mirrored systems--sort of
non-realtime pseudo raid that's refreshed across a gigabit net.
>For the OP FWIW I have noticed greater infant mortality, higher
>incidence of motors which turn noisy, and higher likelihood of
>defective drives which allow wild writes as well as just general
>inconsistency across the ES & PS product line with Maxtor drives over
>the last decade as opposed to most other brands.
As much as I hate to say it, I've probably had more trouble with
Maxtor as well. I was archiving on 250g in USB 2 cases. The
surprising thing is that most have been 5400rpm, which I was
deliberately spec'ing to keep heat down. I like Maxtor the
company, but I'm moving toward Seagate until I figure out
what's up with the Maxtor failures.
I believe Seagate's warranty is 5 years. That shows some
confidence in their product. I feel lucky if I get 5 years out
of a WD or Max.
As for 10 years ago, I was probably using IBM drives. Then
they went downhill and I wouldn't touch them. There's still
some stigma attached for me, even tho they're Hitachi now.
WD has had firmware problems (1993 we had to reflash tons
of drives that were spontaneously powering down). Their tech
support initially denied problems, but later issued new BIOS.
Not a confidence-booster. And I've found their tech support
relatively rude, especially in regard to their own design flaws.
>There is an interesting somewhat anecdotal project over at
>storagereview.com- their reliability database. I try to mention it
>when I can because the more ppl that contribute the better chance we
>have of it yielding accurate results.
Thanks for the reference.
| |
|
| On Sat, 09 Apr 2005 05:02:58 -0000,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>There are persistent rumors around that the lubrication in the
>bearings (both spindle and actuator) will do something nasty like turn
>into tar, if you store a drive for a long time without running it. So
>relying on non-spinning drives for long-term archive seems dangerous.
That was known as 'stiction.' Some drives, like Connor, made in
someone's garage, used to have that problem consistently. (The
lady with the oil can chewed gum?) When UPS *fixes* your
drives enroute there's a problem.
For Connor drives we used to tap the top of the drive to free the
spindle. Later it was a sharp knock. I fixed a stubborn one by
slamming it on a table top. When that didn't work I kicked it
about 15 feet. That didn't work but it felt like the right thing to do
at the time.
>At least not without asking the manufacturer of the drive whether they
>support this mode.
Do you support stiction? Connor: Why yes, we do, sir!
>For my stuff at home, I have two disk copies of everything that's in
>use, plus one copy of all the archival stuff on a single disk. So far
>increasing disk capacity has made using a single disk as an archive
>viable (it is currently a 200GB ATA disk). Plus one extra copy of
>everything (most importantly the archive) on writeable CDs.
I used to use magneto-optical for backup. A bad investment, as
the capacities were outdistanced by drive capacities within a few
months. MO was also much more expensive than the drive that it
was backing up. Very persistant storage though, when you could
get the drives themselves to work.
Some things still go to M.O. Recording studios often use them for
master track backup. I bet the data on my M.O. disks will far outlast
my oxide, CD, or DVD backups, but the last gen of MO that I dealt
with was only 4gigs per disk. Fine for certain things but not for
general backup.
For large volumes of data, I guess the best strategy is to keep it in
at least two places and to constantly get new drives and refresh it.
_R
| |
| Rita Ä Berkowitz 2005-04-09, 8:46 pm |
| _R wrote:
> I don't put myself in the 'SCSI-only, or server' school, but I have
> terabytes of storage (I would be broke if it were SCSI). Much of it
> is for archive purposes. Some of it is on mirrored systems--sort of
> non-realtime pseudo raid that's refreshed across a gigabit net.
Again, you get what you pay for. Most reasonable and sensible people value
their data and time when it comes to disaster prevention and recovery.
> As much as I hate to say it, I've probably had more trouble with
> Maxtor as well. I was archiving on 250g in USB 2 cases. The
> surprising thing is that most have been 5400rpm, which I was
> deliberately spec'ing to keep heat down. I like Maxtor the
> company, but I'm moving toward Seagate until I figure out
> what's up with the Maxtor failures.
Using anything other than an external SCSI solution is pure nonsense!
> I believe Seagate's warranty is 5 years. That shows some
> confidence in their product. I feel lucky if I get 5 years out
> of a WD or Max.
Using anything other than Seagate is totally and utterly foolish!
> As for 10 years ago, I was probably using IBM drives. Then
> they went downhill and I wouldn't touch them. There's still
> some stigma attached for me, even tho they're Hitachi now.
IBM drives were never on top of the hill.
> WD has had firmware problems (1993 we had to reflash tons
> of drives that were spontaneously powering down). Their tech
> support initially denied problems, but later issued new BIOS.
> Not a confidence-booster. And I've found their tech support
> relatively rude, especially in regard to their own design flaws.
Again, using anything other than Seagate is totally and utterly foolish when
you consider the 5-year warranty and on-line RMA process.
[vbcol=seagreen]
Storagereview.com can be a valuable site when you learn how to differentiate
between fact and their biased bullshit.
Rita
| |
|
| flux wrote:
>
> This statement is just fiction. Drives today are roughly the same in reliability, marketing claims notwithstanding.
And drive tech is finicky and rapidly changing, which means you can have
a lot of 'duds' from an otherwise reputable brand. You only discover
this after a year or so, as resin attcks wires, oil evaporates, bearings
crash, heads overheat, or whatnot.
Thomas
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-10, 2:45 am |
| In article <support-BC6D15.02165209042005@news.verizon.net>,
flux <support@fluxsoft.com> wrote:
>In article <1112979104.459192@smirk>,
> _firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>
>Cheap yes, high capacity yes.
>Slow no, unreliable no.
Buy one of each, open them up. In the ES (Enterprise) drive, you'll
find small platters (about 1.5 to 2 inches), spinning at 10K (are
those still available commercially?) or mostly 15K RPM. In the DT
(desktop) drive, you'll find large platters, spinning at 7200 or fewer
RPM. Then look at the actuator magnets. The ones in the ES drive are
much larger, meaning the actuator can move much faster. Then look at
how the platter cavity is sealed against outside air: ES drives are
near perfectly sealed against dust. Then count the air filters: ES
drive have way more.
Now buy a few more drives, and put a few oscilloscopes on them, and
watch what happens while they run. For example, put a scope on the
servo signals (real professionals do this in software of the servo
processor, amateurs have to scope the actuator drive signals), and
watch how accurately the drive is servoing. Now start vibrating the
drive (simulating the effect of the cooling fan in the computer, or
the second drive or CD-ROM next to the drive), and watch how well the
servo tracks the vibration. You'll find that the DT drive is servoing
rather crudely, and can' handle vibration well (why: servo processor
much less powerful). Then start doing IO while watching the servo.
You'll most likely find that the DT drive will actually stop servoing
while writing, with the head flying blind (why: servo processor has to
do double duty running the data path, while an ES drive has separate
hardware for that).
Look at the PC boards of the two boards, and count the chips. Measure
the Flash-ROM capacity of each drive (gives a crude indication of the
complexity of the firmware of each drive, that is the software
development cost that went into the drive). Get the wholesale price
of all the chips (you'll have to estimate the ASICs by measuring die
area). You'll find a significant difference here.
Replace the scope on the actuator with a storage scope, and measure
how long the actuator really needs to settle on a track after a move
(you can also do this in software, but that's actually tricky
business). You'll find that ES actuators move about twice as fast
(and suck way more power than DT drives while doing that, there is no
free lunch). Actually: drives generally use much more power when busy
(writing and in particular seeking), so take a bandsaw to a few
drives, and measure how thick the aluminum and copper structures that
conduct heat away to the frame are; after all, you want to keep the
drive cool (heat is the enemy of just about everything in the world).
If you want to be neat about it, open the drive, and put a dozen
thermocouples in strategic locations, button it back up, and run it
for a while.
Now throw a nasty workload at the drive, with a few dozen outstanding
IOs at the same time (classic ATA drives can't do that at all, modern
ATA drives can do it in principle, but because that capability is not
used by Windows device drivers, it is implemented in a haphazard
fashion). Watch the order in which the drive actually executes the
IOs, and how many IOs it can queue. You'll find that under overload
conditions, the ES drive acquits itself fairly well, using a pretty
sane queue management algorithm. On the DT drive, it's anyones guess.
Doing these tests will require a well-equipped lab, a few dozen
drives, and a few weeks of time (if you are good at this kind of
stuff). Or friends in the business that have done these tests, and in
some cases published them (I pointed to the Dykes/Anderson/Riedel
paper in a previous post).
>
>Let's have the names.
Please ask your employer to get them through official channels. Don't
forget to sign the requisite NDA agreements, and you might have to pay
a little license fee here or there for information like this. Don't
worry, if you use a few hundred thousand drives per year, the extra
cost will not be significant.
>
>Are there any such drives any more?
Why are most ES drives sold with 3- and 5-year warranties, and DT
drives with much shorter warranties (1 year at many manufacturers)?
Coincidence?
>I'm not sure I believe you.
Too bad. No skin off my avocado.
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
|
| In article <VYGdnb2rQ51GTMrfRVn-qA@metrocastcablevision.com>,
Bill Todd <billtodd@metrocast.net> wrote:
> Fact: ATA drives max out at 7200 rpm, while SCSI/FC drives max out at
> 15,000 rpm. That gives SCSI drives a 2+:1 advantage in rotational
> latency right off the bat.
That doesn't necessary mean they are faster. In fact, they can even be slower.
> As for reliability, I'll let you read the paper that you previously
> neglected to: it's quite thorough in its assessment of the differences.
Apparently not.
| |
|
| In article <1113111357.989401@smirk>,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
> In article <support-BC6D15.02165209042005@news.verizon.net>,
> flux <support@fluxsoft.com> wrote:
>
[vbcol=seagreen]
> Buy one of each, open them up. In the ES (Enterprise) drive, you'll
> find small platters (about 1.5 to 2 inches), spinning at 10K (are
> those still available commercially?) or mostly 15K RPM. In the DT
> (desktop) drive, you'll find large platters, spinning at 7200 or fewer
> RPM. Then look at the actuator magnets. The ones in the ES drive are
> much larger, meaning the actuator can move much faster. Then look at
> how the platter cavity is sealed against outside air: ES drives are
> near perfectly sealed against dust. Then count the air filters: ES
> drive have way more.
>
> Now buy a few more drives, and put a few oscilloscopes on them, and
> watch what happens while they run. For example, put a scope on the
> servo signals (real professionals do this in software of the servo
> processor, amateurs have to scope the actuator drive signals), and
> watch how accurately the drive is servoing. Now start vibrating the
> drive (simulating the effect of the cooling fan in the computer, or
> the second drive or CD-ROM next to the drive), and watch how well the
> servo tracks the vibration. You'll find that the DT drive is servoing
> rather crudely, and can' handle vibration well (why: servo processor
> much less powerful). Then start doing IO while watching the servo.
> You'll most likely find that the DT drive will actually stop servoing
> while writing, with the head flying blind (why: servo processor has to
> do double duty running the data path, while an ES drive has separate
> hardware for that).
>
> Look at the PC boards of the two boards, and count the chips. Measure
> the Flash-ROM capacity of each drive (gives a crude indication of the
> complexity of the firmware of each drive, that is the software
> development cost that went into the drive). Get the wholesale price
> of all the chips (you'll have to estimate the ASICs by measuring die
> area). You'll find a significant difference here.
>
> Replace the scope on the actuator with a storage scope, and measure
> how long the actuator really needs to settle on a track after a move
> (you can also do this in software, but that's actually tricky
> business). You'll find that ES actuators move about twice as fast
> (and suck way more power than DT drives while doing that, there is no
> free lunch). Actually: drives generally use much more power when busy
> (writing and in particular seeking), so take a bandsaw to a few
> drives, and measure how thick the aluminum and copper structures that
> conduct heat away to the frame are; after all, you want to keep the
> drive cool (heat is the enemy of just about everything in the world).
> If you want to be neat about it, open the drive, and put a dozen
> thermocouples in strategic locations, button it back up, and run it
> for a while.
>
> Now throw a nasty workload at the drive, with a few dozen outstanding
> IOs at the same time (classic ATA drives can't do that at all, modern
> ATA drives can do it in principle, but because that capability is not
> used by Windows device drivers, it is implemented in a haphazard
> fashion). Watch the order in which the drive actually executes the
> IOs, and how many IOs it can queue. You'll find that under overload
> conditions, the ES drive acquits itself fairly well, using a pretty
> sane queue management algorithm. On the DT drive, it's anyones guess.
>
> Doing these tests will require a well-equipped lab, a few dozen
> drives, and a few weeks of time (if you are good at this kind of
> stuff). Or friends in the business that have done these tests, and in
> some cases published them (I pointed to the Dykes/Anderson/Riedel
> paper in a previous post).
In other words, it's just as I said:
Cheap yes, high capacity yes.
Slow no, unreliable no.
>
> Please ask your employer to get them through official channels. Don't
Let's have the names.
>
> Why are most ES drives sold with 3- and 5-year warranties, and DT
> drives with much shorter warranties (1 year at many manufacturers)?
Are you sure about that?
| |
|
| In article <nYadnXH2s5YgU8rfRVn-vg@metrocastcablevision.com>,
Bill Todd <billtodd@metrocast.net> wrote:
> flux wrote:
>
> ...
>
>
> Then you might benefit from actually looking at the material he cited
> rather than pulling opinions directly out of your XXX.
I don't believe him. Why should I believe them?
| |
| Bill Todd 2005-04-10, 5:46 pm |
| flux wrote:
> In article <nYadnXH2s5YgU8rfRVn-vg@metrocastcablevision.com>,
> Bill Todd <billtodd@metrocast.net> wrote:
>
>
>
>
> I don't believe him. Why should I believe them?
Because you're not a complete idiot, perhaps?
On the other hand, faith-based beliefs that admit to no factual
refutation are pretty popular right now, so...
- bill
| |
| Bill Todd 2005-04-10, 5:46 pm |
| flux wrote:
> In article <VYGdnb2rQ51GTMrfRVn-qA@metrocastcablevision.com>,
> Bill Todd <billtodd@metrocast.net> wrote:
>
>
>
>
> That doesn't necessary mean they are faster. In fact, they can even be slower.
No, you moron: it *does* necessarily mean they are faster, for at least
that particular metric. And if you check out sequential bit rates from
the platters, you'll find that current SCSI/FC drives are faster there
as well.
>
>
>
>
> Apparently not.
Certainly sufficiently thorough to put your mistaken impressions to rest.
- bill
| |
| Curious George 2005-04-10, 5:46 pm |
| On Sat, 09 Apr 2005 09:29:45 -0400, _R <_R@nomail.org> wrote:
>Are there other storage groups? I only know of this one.
comp.sys.ibm.pc.storage
comp.mac.hardware.storage
fido.ger.storage
tw.bbs.comp.hardware.storage
alt.comp.hardware.superdisk
microsoft.public.win98.disks.general
etc...
The topic comes up in every hardware or computer oriented forum,
whether web board, bbs, irc, usenet, etc and the dialog is usually
very predictable
>WD has had firmware problems (1993 we had to reflash tons
>of drives that were spontaneously powering down). Their tech
>support initially denied problems, but later issued new BIOS.
>Not a confidence-booster. And I've found their tech support
>relatively rude, especially in regard to their own design flaws.
Other WD firmware issues have been observed even more recently. With
ATA the problem can't be flashed out.
| |
| Curious George 2005-04-10, 5:46 pm |
| On Sat, 9 Apr 2005 10:00:45 -0400, "Rita Ä Berkowitz" <ritaberk2O04
@aol.com> wrote:
>_R wrote:
>
>
>Again, you get what you pay for. Most reasonable and sensible people value
>their data and time when it comes to disaster prevention and recovery.
For many smaller organizations or individuals there simply aren't the
funds for terrabytes of scsi storage while ata may at least appear
doable.
Being scsi isn't the whole issue. Let's say someone bought a batch of
used, old FC or SCSI Seagates' off eBay. They're not magically
protected by invulnerable bulletproof drives.
>
>Using anything other than an external SCSI solution is pure nonsense!
not if you require compatibility with other ppl's systems. Let's say
you're a consultant hired to fix other ppl's problems or configure on
systems you didn't get a chance to sell them. External SCSI is the
worst idea possible for portable large capacity media.
>
>Using anything other than Seagate is totally and utterly foolish!
I love Rita-speak. Its broken-recordese.
>
>IBM drives were never on top of the hill.
Even when they were inventing the technology ;)
>
>Again, using anything other than Seagate is totally and utterly foolish when
>you consider the 5-year warranty and on-line RMA process.
Come one. Everybody's refurbs are crap they're trying to pawn off on
an unlucky slob. Testing & remanufacturing is always cursory at best.
Otherwise it would be cheaper to just take one off the cookie-cutter
assembly line. Seagate is no exception. That greatly diminishes the
usefulness of a 5 yr warranty.
>
>Storagereview.com can be a valuable site when you learn how to differentiate
>between fact and their biased bullshit.
That applies to every information source. I called it "anecdotal" for
a reason. Unknown credentials, unknown financial relationships, and
unknown axes to grind are a part of every opinion stated as fact.
But you know all about that don't you Rita?
| |
| Curious George 2005-04-10, 5:46 pm |
| On Sat, 09 Apr 2005 06:22:05 GMT, flux <support@fluxsoft.com> wrote:
>I don't know how much of it is true,
we know
>but I do know that ATA drives are a
>lot faster and a lot more reliable than some here think they are.
Yeah the STR is good & so is the idea of lower overhead but throw a
couple of disk intensive tasks at an ATA disk or array & it chokes.
ATA is fast enough for light duty use but chokes in a stressfully
environment. ATA RAID is often downright pathetic.
There is a difference between something basically working outside of a
few occasional, recoverable quirks over a lifetime that seems mostly
OK 'cause there isn't really much important on it anyway and doesn't
get used all that much & something doing exactly what it's supposed to
do all the time and exceeding expected life while being able to
sustain much harder use (longer head use).
Unfortunately, while better drives tend to be scsi or fibre that's not
to say every one is and has been bulletproof. I'll take a reliable
ATA over a crummy older SCSI drive any time. My druthers is with more
solid scsi or FC models any day of the week & twice on sundays.
I know I'm not going to change your mind. I've seen your kind of
predictable responses too often. They always come from either troll
or inexperience.
| |
| K.P.King 2005-04-10, 5:46 pm |
| Curious George wrote:
> On Sat, 09 Apr 2005 06:22:05 GMT, flux <support@fluxsoft.com> wrote:
>
>
>
>
> we know
>
>
>
>
> Yeah the STR is good & so is the idea of lower overhead but throw a
> couple of disk intensive tasks at an ATA disk or array & it chokes.
> ATA is fast enough for light duty use but chokes in a stressfully
> environment. ATA RAID is often downright pathetic.
>
> There is a difference between something basically working outside of a
> few occasional, recoverable quirks over a lifetime that seems mostly
> OK 'cause there isn't really much important on it anyway and doesn't
> get used all that much & something doing exactly what it's supposed to
> do all the time and exceeding expected life while being able to
> sustain much harder use (longer head use).
>
> Unfortunately, while better drives tend to be scsi or fibre that's not
> to say every one is and has been bulletproof. I'll take a reliable
> ATA over a crummy older SCSI drive any time. My druthers is with more
> solid scsi or FC models any day of the week & twice on sundays.
>
> I know I'm not going to change your mind. I've seen your kind of
> predictable responses too often. They always come from either troll
> or inexperience.
Whoops - apologies for the earlier posts which were sent in error.
What I can add to this discussion is that the enterprise vs. desktop
differences are tangible and not simply a case of sampling yields, as
has been the case with speed ratings for semi-conductors.
Bandwidth and I/O issues aside, on the topic of reliability, the key
issue is the MTBF and bit error rates, which are broadly similar.
However, my understanding are the tests which lead to these figures are
different.
It is expected that in some applications SCSI/FC drives are
running 24/7 at high intensity, therefore the test figures quote for this.
The profile of general desktop usage is completely different, and the
tests are geared for the expected pattern of usage in this market, i.e.
sporadic usage, probably more power cycles, etc.
Also, the enterprise stuff tends to have more comprehensive fault
analysis, i.e. SMART and the like, which can be of use to the RAID
vendor in pre-emptively signaling a failing drive. General operating
system support for this stuff is patchy and in (S)ATA drives will
normally only be of use in a post mortem scenario, or through some
dedicated diagnostic software.
Regards,
K
| |
|
| In article <goGdncCPG9qAy8TfRVn-pQ@metrocastcablevision.com>,
Bill Todd <billtodd@metrocast.net> wrote:
> flux wrote:
>
> No, you moron: it *does* necessarily mean they are faster, for at least
No, it doesn't.
> that particular metric. And if you check out sequential bit rates from
> the platters, you'll find that current SCSI/FC drives are faster there
> as well.
>
>
> Certainly sufficiently thorough to put your mistaken impressions to rest.
It wasn't expressing an impression.
| |
|
| In article <nYadnXH2s5YgU8rfRVn-vg@metrocastcablevision.com>,
Bill Todd <billtodd@metrocast.net> wrote:
> flux wrote:
>
> ...
>
>
> Then you might benefit from actually looking at the material he cited
And that changes facts how?
| |
|
| In article <3goi51d34ime5a4lrrm7qvhea9gehkba4h@4ax.com>,
Curious George <cg@email.net> wrote:
> On Sat, 09 Apr 2005 06:22:05 GMT, flux <support@fluxsoft.com> wrote:
>
>
> we know
>
>
> Yeah the STR is good & so is the idea of lower overhead but throw a
> couple of disk intensive tasks at an ATA disk or array & it chokes.
SATA certainly doesn't share this problem.
> ATA is fast enough for light duty use but chokes in a stressfully
> environment. ATA RAID is often downright pathetic.
SATA doesn't share this problem either. Is it that different or could it be that ATA is not as bad as you think?
> Unfortunately, while better drives tend to be scsi or fibre that's not
> to say every one is and has been bulletproof. I'll take a reliable
> ATA over a crummy older SCSI drive any time. My druthers is with more
I'm confused. Some people seem to be arguing you can't say reliable and ATA in the same sentence.
> solid scsi or FC models any day of the week & twice on sundays.
>
> I know I'm not going to change your mind. I've seen your kind of
> predictable responses too often. They always come from either troll
> or inexperience.
Or they happen to be true.
| |
|
| In article <Vxh6e.16880$il.1195@newsfe5-win.ntli.net>,
"K.P.King" <k.p.king@ntlworld.com> wrote:
> It is expected that in some applications SCSI/FC drives are
> running 24/7 at high intensity, therefore the test figures quote for this.
>
> The profile of general desktop usage is completely different, and the
> tests are geared for the expected pattern of usage in this market, i.e.
> sporadic usage, probably more power cycles, etc.
>
Actually, it's not. Desktop drives are in use 24/7. Just check out video
recorders like Tivo. These things record video 24/7. AFAIK, they are
just ordinary ATA drives.
Another thing relates to desktop computers. If there are to be backed
up, that usually happens at night and therefore they must be left
running. So I suspect it's actually unusual for desktop computers--at
least those in a workplace--not to be running 24/7.
On the other hand, I wonder if takes more out of a drive to powered up
and then powered down again and again. Seems like 24/7 usage is less
usage, strangely enough.
> Also, the enterprise stuff tends to have more comprehensive fault
> analysis, i.e. SMART and the like, which can be of use to the RAID
> vendor in pre-emptively signaling a failing drive. General operating
> system support for this stuff is patchy
What operating system has builtin support?
| |
| Matthias Buelow 2005-04-12, 7:46 am |
| flux <support@fluxsoft.com> writes:
> Actually, it's not. Desktop drives are in use 24/7. Just check out video
> recorders like Tivo. These things record video 24/7. AFAIK, they are
> just ordinary ATA drives.
The issue is not the platters rotating but head move (constant
seeking). Put an ordinary ATA disk in a busy newsserver and watch it
explode.
mkb.
| |
| Michael Haardt 2005-04-12, 5:47 pm |
| Matthias Buelow <mkb@incubus.de> writes:
> The issue is not the platters rotating but head move (constant
> seeking). Put an ordinary ATA disk in a busy newsserver and watch it
> explode.
Good thing I did not spend time watching, as nothing exploded since
I run ordinary (WD400, if that counts) drives in my newsfeeders. ;-)
Sure, there is a bad drive every now and then, but drive failures
happen with expensive SCSI drives, too. The rate does not worry me.
I found that I just don't get much more than 130-140 I/O transactions
per drive on average, but I had that issue with the expensive SCSI drives
before, too. More spindles take care of that for my applications.
Back to the original subject: I had bad series with most manufacturers
at some time over the years and I am currently using WD, because their
RMA handling works without trouble and because their drives are pretty
fast, so I have no reason to switch at this time.
Michael
| |
| Curious George 2005-04-13, 2:46 am |
| On Tue, 12 Apr 2005 04:09:24 GMT, flux <support@fluxsoft.com> wrote:
>In article <3goi51d34ime5a4lrrm7qvhea9gehkba4h@4ax.com>,
> Curious George <cg@email.net> wrote:
>
>
>SATA certainly doesn't share this problem.
sorry but PS drives are the same regardless of interface, unless you
mean raptors specifically. SATA kills the lower overhead advantage
PATA had over SCSI.
>
>SATA doesn't share this problem either. Is it that different or could it be that ATA is not as bad as you think?
I'm not taking about sharing the bus, just raw disk performance with
more complex/demanding usage than typical desktop usage patterns. If
you're thinking of raptors specifically state that. Raptors are not
typical SATA drives. They are closer to apples & apples comparison
but then the price is similar also to scsi while at the same time
being newer/less mature/less proven track record.
Right now I'm using/evaluating/testing Seagate 7200.8's in arrays.
Even though the synthetic benchmarks are basically lining up to what
they're supposed to be it still chokes very easily. Very
disappointing...
I'm sorry but I don't see the sense of raptors and a good 3ware card
or whatever. I don't care whether performance & reliability is
competitive or not. 10K SCSI makes more sense to me. It's more
mature, more flexible, better supported, has a longer track record,
etc. and costs about the same. 1st gen 10k sata compared to 6th or
7th gen 10k scsi. Come on.
>
>I'm confused. Some people seem to be arguing you can't say reliable and ATA in the same sentence.
It's not either/or. There are bum models of every interface.
Although the best SCSI tends to beat the best ATA reliability wise
hands down, but then you pay for that. But if you go through more ATA
drives in the lifetime of a machine, cluster, etc, (even a small
amount more) you pay for that too- it ultimately costs much more
despite the lower up front costs like parts.
In many businesses even marginal increases in reliability are a big
deal because of the massive costs of support, maintenance and of
service interruption. One's attitude depends on individual tolerance
of risk & fiddling around. ATA doesn't have to be totally unstable
garbage to be/seem unsuitable to many ppl & environments.
>
>Or they happen to be true.
very cute.
| |
|
| In article <86zmw41eqn.fsf@drjekyll.mkbuelow.net>,
Matthias Buelow <mkb@incubus.de> wrote:
> flux <support@fluxsoft.com> writes:
>
>
> The issue is not the platters rotating but head move (constant
> seeking). Put an ordinary ATA disk in a busy newsserver and watch it
> explode.
That's probably just a fraction of the work a Tivo machine requires of it.
| |
| Paul Rubin 2005-04-13, 2:46 am |
| Matthias Buelow <mkb@incubus.de> writes:
> The issue is not the platters rotating but head move (constant
> seeking). Put an ordinary ATA disk in a busy newsserver and watch it
> explode.
That makes me wonder how fast cheap ATA drives explode when installed
in busy RAID arrays. I wonder if the drive vendors take a beating on
warranty service on drives that have been in RAID's.
| |
|
| In article <uq7p51584na6ou9cmsr5c1eqckp53kdpuu@4ax.com>,
Curious George <cg@email.net> wrote:
>
> I'm not taking about sharing the bus, just raw disk performance with
> more complex/demanding usage than typical desktop usage patterns. If
> you're thinking of raptors specifically state that. Raptors are not
> typical SATA drives. They are closer to apples & apples comparison
> but then the price is similar also to scsi while at the same time
> being newer/less mature/less proven track record.
>
> Right now I'm using/evaluating/testing Seagate 7200.8's in arrays.
> Even though the synthetic benchmarks are basically lining up to what
> they're supposed to be it still chokes very easily. Very
> disappointing...
What is the performance and what does it mean to choke? Are you actually
saying the hard drive starts slowing down and even stops outright,
crashes?
> I'm sorry but I don't see the sense of raptors and a good 3ware card
> or whatever. I don't care whether performance & reliability is
> competitive or not. 10K SCSI makes more sense to me. It's more
> mature, more flexible, better supported, has a longer track record,
I don't see how it's more flexible. In fact, it's less flexible. SCSI
has to deal ID numbers and termination. This is irrelevant to SATA. SATA
also has slim cables.
What does it mean to better supported? Seems to me they are the same
manufacturers with the same RMA procedures.
> etc. and costs about the same.
Costs the same? Where do you get SCSI drives this cheap?
> 1st gen 10k sata compared to 6th or
> 7th gen 10k scsi. Come on.
Isn't SATA supposed to be the successor to ATA-6? That would make SATA
7th generation? Now the technology is different from ATA somewhat, so
it's not exactly 7th generation, but it's hardly a first generation
drive. That would be old Winchesters, no?
> hands down, but then you pay for that. But if you go through more ATA
> drives in the lifetime of a machine, cluster, etc, (even a small
The thing is I'm not sure that it actually happens that way. What I see
is that the actual electronics of the computer: motherboards, RAM, CPU
break down as well. I would say motherboard are most likely component
to fail and power supplies the least. Apparently, this is the opposite
of the accepted wisdom.
Also, when a drive goes down, it could be an electronics failure. How's
that different from any other electronic component failure or is anyone
arguing that the ATA drives somehow get less reliable versions of these
components than other parts of a computer?
| |
| Curious George 2005-04-13, 2:46 am |
| On Wed, 13 Apr 2005 05:25:03 GMT, flux <support@fluxsoft.com> wrote:
>In article <86zmw41eqn.fsf@drjekyll.mkbuelow.net>,
> Matthias Buelow <mkb@incubus.de> wrote:
>
>
>That's probably just a fraction of the work a Tivo machine requires of it.
???
| |
| Curious George 2005-04-13, 2:46 am |
| On Wed, 13 Apr 2005 05:39:18 GMT, flux <support@fluxsoft.com> wrote:
>In article <uq7p51584na6ou9cmsr5c1eqckp53kdpuu@4ax.com>,
> Curious George <cg@email.net> wrote:
>
>
>What is the performance and what does it mean to choke? Are you actually
>saying the hard drive starts slowing down and even stops outright,
>crashes?
Performance drops off severely with multiple IO transactions/
demanding head use. More severely than I expected at least (my bad I
guess)
>
>I don't see how it's more flexible. In fact, it's less flexible. SCSI
>has to deal ID numbers and termination. This is irrelevant to SATA. SATA
>also has slim cables.
Longer cable lengths, better options for external esp more robust
boxes. Controllers tend to be better. Better EC/checksumming. ID
numbers & termination are far from complicated. Most sata doesn't
allow remote or delay start and led comes from the controller instead
of the drive. ALL PITA for integration. Many PPL also complain about
cable retention & other PITA issues. Most SATA drives lack normal
jumper options of scsi which can be helpful.
>What does it mean to better supported? Seems to me they are the same
>manufacturers with the same RMA procedures.
mostly via software management & diagnostics, modepages, etc.
Sometimes also vendor, etc.
SATA will be better along these lines, but not yet really.
>
>Costs the same?
For example:
Raptor vs, say cheetah 10K.6
3ware9500 vs LSI Megaraid
>Where do you get SCSI drives this cheap?
anywhere:
www.pricegrabber.com
>
>Isn't SATA supposed to be the successor to ATA-6? That would make SATA
>7th generation? Now the technology is different from ATA somewhat, so
>it's not exactly 7th generation, but it's hardly a first generation
>drive. That would be old Winchesters, no?
Raptors are WD's first attempt at ES in a long time. There past scsi
drives left a lot to be desired IMHO. Raptor is the first attempt of
anyone at 10K ES SATA.
The interface was radically redesigned for SATA. SATA 1 essentially
specified a SATA->PATA bridge. True SATA is very new even though it
builds on the older protocols of scsi & ata.
Don't forget the ATA/ATAPI standards mean little to real world
products. There's a whole mess of stuff that makes it look good on
paper but which isn't implemented in products, or in some cases not
implemented well.
>
>The thing is I'm not sure that it actually happens that way. What I see
>is that the actual electronics of the computer: motherboards, RAM, CPU
>break down as well. I would say motherboard are most likely component
>to fail and power supplies the least. Apparently, this is the opposite
>of the accepted wisdom.
Not true, unless you are accustomed to buying cheap crap mobos and
horrible cases. The disks should normally be the first to go or at
least hickup followed by the fans & maybe PSU. It's the moving parts
that fail first. IC's, etc. last a long time in a proper environment,
even under full load. Embedded systems are used in harsh environments
for a reason.
>Also, when a drive goes down, it could be an electronics failure. How's
>that different from any other electronic component failure or is anyone
>arguing that the ATA drives somehow get less reliable versions of these
>components than other parts of a computer?
firstname_@lr_dot_los-gatos_dot_ca.us posted a pretty good dissection
of the build differences. You should read that as well as the other
article cited.
| |
| Matthias Buelow 2005-04-13, 7:46 am |
| flux <support@fluxsoft.com> writes:
>
> That's probably just a fraction of the work a Tivo machine requires of it.
Care to explain how you reach this conclusion?
mkb.
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-13, 5:46 pm |
| In article <support-19658E.01250313042005@news.verizon.net>,
flux <support@fluxsoft.com> wrote:
>In article <86zmw41eqn.fsf@drjekyll.mkbuelow.net>,
> Matthias Buelow <mkb@incubus.de> wrote:
>
>
>That's probably just a fraction of the work a Tivo machine requires of it.
Can someone who has more patience than I do please explain the
difference between sequential and random access patterns? And give a
little lecture on the performance characteristics of disks when
exposed to those access patterns? I would guess that operating a Tivo
(at maybe a few MB/s, nearly completely sequential, large IOs) barely
stresses a slow ATA disk, whereas news servers tend to be disk-limited
with random IO and short IOs (you add disks until the disks are barely
capable of keeping up, so the disks are always close to being
overloaded).
I would like to add the following: If you watch a news server, you'll
find that it is quite busy 24x7. Matter-of-fact, it isn't clear that
in these days of global news distribution there is a quiet time at
night. This is also true of many corporate servers: During the day,
they are running transaction processing and web-driven workloads;
during the night they are running data mining, analytics, and backup.
The access patterns are different, but they tend to be busy all the
time. Why? If they were not busy some time of the day, they are
underutilized, and the workloads that can be rescheduled will be moved
to the underutilized time.
In contrast: In most households, nodoby will be watching on the Tivo
between midnight and 6AM, and little during the day (while everyone is
at work). Also, there is little stuff on TV worth recording in the
dark of the night and the middle of the day, so a Tivo is likely idle
about 50% of the time.
Note that I said "idle", not powered down. This is about actuator,
not about spindle.
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
| _firstname_@lr_dot_los-gatos_dot_ca.us 2005-04-13, 5:46 pm |
| In article <uq7p51584na6ou9cmsr5c1eqckp53kdpuu@4ax.com>,
Curious George <cg@email.net> wrote:
>I'm not taking about sharing the bus, just raw disk performance with
>more complex/demanding usage than typical desktop usage patterns. If
>you're thinking of raptors specifically state that. Raptors are not
>typical SATA drives. They are closer to apples & apples comparison
>but then the price is similar also to scsi while at the same time
>being newer/less mature/less proven track record.
That's exactly the point of the Anderson/Dykes/Riedel paper: There are
different kind of drives (namely enterprise and personal/desktop, with
some halfway crossover models). Then there are different interfaces
(for example ATA, SATA, SCSI, FC). There was a very strong
correlation between those two characteristics a few years ago: Cheap,
slow, large capacity, unreliable drives tended to be ATA, while
expensive, fast, smaller capacity, reliable drives tended to be FC.
The arrival of purported enterprise-grade ATA disks and of SATA disks
has muddled this strong correlation. But the correlation in the
marketplace does not HAVE to be true. Other than the lack of demand,
there is very little that prevents disk manufacturers from making
ultra-high-end enterprise drives with ATA or SATA interfaces, and
cheap consumer drives with FC interfaces. Clearly, there is logical
reasons for the lack of demand.
>I'm sorry but I don't see the sense of raptors and a good 3ware card
>or whatever. I don't care whether performance & reliability is
>competitive or not. 10K SCSI makes more sense to me. It's more
>mature, more flexible, better supported, has a longer track record,
>etc. and costs about the same. 1st gen 10k sata compared to 6th or
>7th gen 10k scsi. Come on.
Even though I very much agree with you, we have to admit that there
are rare exceptions where storage farms built out of inexpensive RAID
cards with cheap consumer-grade disks make a lot of sense. These tend
to be environments that are very large (so they can amortize the extra
management overhead of having to regularly replace failed disks),
require a heck of a lot of storage at low IO intensities, and can
tolerate and manage data loss. I know several examples of disks farms
that use thousands or tenthousands of ATA disks this way, typically
with exactly the 3ware cards you mentioned. But the bulk of the
industry will be using high-reliability disk arrays, constructed and
supported by high-end vendors, because the significantly higher
purchase and support cost is worth the saving in hassle, data loss,
and systems management.
>In many businesses even marginal increases in reliability are a big
>deal because of the massive costs of support, maintenance and of
>service interruption. One's attitude depends on individual tolerance
>of risk & fiddling around. ATA doesn't have to be totally unstable
>garbage to be/seem unsuitable to many ppl & environments.
Again, as much as I agree in general, there are examples where support
and systems management are not relevant as cost factors. Hobbyists
are one example, there are others. In those environments, using
hard-to-manage components that need to be babied to prevent data loss
and need to be assembled from store-bought components is very
sensible. If I had a little more spare time, I might set up an ATA
RAID system at home with used WD drives and a 3ware card (if I can get
the hardware for cheap at a surplus store, I'm notoriously stingy).
In the meantime, I feel more comfortable with my 10K RPM SCSI drive
for the data I really care about. But I wouldn't even dream of
setting up such a system in my job, or for a serious customer that is
paying for storage.
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
| Ralph Becker-Szendy 2005-04-13, 5:46 pm |
| In article <fscp51hllq36ihajkrpv44dmmd7npbl0oh@4ax.com>,
Curious George <cg@email.net> wrote:
>Not true, unless you are accustomed to buying cheap crap mobos and
>horrible cases. The disks should normally be the first to go or at
>least hickup followed by the fans & maybe PSU. It's the moving parts
>that fail first. IC's, etc. last a long time in a proper environment,
>even under full load. Embedded systems are used in harsh environments
>for a reason.
Furthermore, on real computers, power supplies and fans are all
redundant, and can be hot-swapped. Look at the back of a good
rackmount computer sometime (in particular when looking at an
enterprise-style Unix machine, like a HP-UX or AIX box).
In contrast, disks are by their nature difficult to hot-swap, because
when you put the new disk in, it doesn't have any useful bits on it.
This is where RAID comes in. But even that's pretty difficult. PATA
was not designed for hot-swap; the fact that some disk enclosures and
the 3ware cards can do it at all is a bit of a miracle. At least SATA
is designed for hot swap. All SCSI and FC drives sold today are
hot-swappable (that's why many of the SCSI drives are sold with the
80-pin connector that integrated data and power in one connector).
But even after you hot-swap, the RAID controller has to do a lot of
work to put the useful bits back on the drive. During this
reconstruction period, the other drive(s) in the RAID group are going
to be heavily loaded, often to the detriment of the foreground
workload. Also, while the drive is removed, and while the new drive
is being rebuilt onto, you are running with no redundancy (less
redundancy if you were running with a RAID setup that tolerates
multiple failures, but those are still exceedingly rare outside of
high-end enterprise disk arrays).
In summary, the disks are and remain the least reliably component of a
serious computer system.
Now, obviously by buying crap components, you can make arbitrarily bad
systems. If I only used motherboards from the grab bin at the surplus
store, and power supplies that are cheap because they failed the
burn-in test at the manufacturer, my disk drives might actually look
good in contrast to this crap. Nobody who cares about their computers
would build a system that way (masochists excepted).
--
The address in the header is invalid for obvious reasons. Please
reconstruct the address from the information below (look for _).
Ralph Becker-Szendy _firstname_@lr_dot_los-gatos_dot_ca.us
| |
| Jim Prescott 2005-04-13, 5:46 pm |
| In article <1113415288.189491@smirk>,
<_firstname_@lr_dot_los-gatos_dot_ca.us> wrote:
>In contrast: In most households, nodoby will be watching on the Tivo
>between midnight and 6AM, and little during the day (while everyone is
>at work). Also, there is little stuff on TV worth recording in the
>dark of the night and the middle of the day, so a Tivo is likely idle
>about 50% of the time.
>Note that I said "idle", not powered down. This is about actuator,
>not about spindle.
Tivo records 24x7. It works harder when you are watching since it is
also reading, but it is always recording.
I'm not stating an opinion on the larger thread here. I just think
Tivo might not be that meaningful in a discussion about disk
reliability. Tivo's disk needs to be quiet, not generate much heat and
have "enough" performance (ie having more than enough performance
doesn't really help it at all). Computers will typically have a very
different set of requirements.
--
Jim Prescott - Computing and Networking Group jgp@seas.rochester.edu
School of Engineering and Applied Sciences, university of Rochester, NY
| |
| Paul Rubin 2005-04-13, 5:46 pm |
| jgp@harn.ceas.rochester.edu (Jim Prescott) writes:
>
> Tivo records 24x7. It works harder when you are watching since it is
> also reading, but it is always recording.
I think even if it's recording 24x7, it's probably not seeking 24x7.
It's likely recording whole tracks contiguously.
| |
|
| In article <1113415288.189491@smirk>,
_firstname_@lr_dot_los-gatos_dot_ca.us wrote:
> exposed to those access patterns? I would guess that operating a Tivo
> (at maybe a few MB/s, nearly completely sequential, large IOs) barely
Your guess is wrong.
> during the night they are running data mining, analytics, and backup.
> The access patterns are different, but they tend to be busy all the
> time. Why? If they were not busy some time of the day, they are
There is no why because there probably isn't much difference.
> In contrast: In most households, nodoby will be watching on the Tivo
> between midnight and 6AM, and little during the day (while everyone is
> at work). Also, there is little stuff on TV worth recording in the
> dark of the night and the middle of the day, so a Tivo is likely idle
> about 50% of the time.
The only time a Tivo is idle if when there is a power failure.
| |
|
| In article <fscp51hllq36ihajkrpv44dmmd7npbl0oh@4ax.com>,
Curious George <cg@email.net> wrote:
>
> Performance drops off severely with multiple IO transactions/
> demanding head use. More severely than I expected at least (my bad I
> guess)
Let's see the data.
>
> Longer cable lengths,
Why do you need long cables?
> boxes. Controllers tend to be better.
What does that mean exactly?
> numbers & termination are far from complicated. Most sata doesn't
> allow remote or delay start and led comes from the controller instead
> of the drive.
Is that really the case?
> ALL PITA for integration. Many PPL also complain about
> cable retention & other PITA issues.
What do you mean by cable retention? You can get locking SATA cables.
Probably a real difference is the number of devices that can be
connected a to a controller.
> Most SATA drives lack normal
> jumper options of scsi which can be helpful.
When do you ever use these jumpers?
>
> mostly via software management & diagnostics, modepages, etc.
I'm not sure. Promise makes SATA RAID boxes with a very nice set of
management tools. And 3Ware is also pretty good.
> SATA will be better along these lines, but not yet really.
>
>
> For example:
> Raptor vs, say cheetah 10K.6
Seagate:
http://www.spacecentersystems.com/c.../products_id/94
37?refsrc=froogle
Raptor:
http://www.tritechcoa.com/product/067848.html
That's a $100 difference. It adds up if you need 50 of them.
> 3ware9500 vs LSI Megaraid
>
This is a reasonable comparison. More devices can be hooked up to the
LSI card.
>
> Not true, unless you are accustomed to buying cheap crap mobos and
> horrible cases.
It really is true! Really.
If I mention Dell, does that mean you tell me they use cheap crap mobos
and horrible cases?
> The disks should normally be the first to go or at
> least hickup followed by the fans & maybe PSU. It's the moving parts
> that fail first. IC's, etc. last a long time in a proper environment,
> even under full load. Embedded systems are used in harsh environments
> for a reason.
I agree that it should be this way and everybody runs around telling
people it's true--you are telling us---and it makes sense and that it
probably *has been* true in the past. But today, things apparently have
changed. Moving parts are as robust if not more so than the electronics.
Perhaps we should consider applying RAID to motherboards, CPUs and
memory.
| |
| Thor Lancelot Simon 2005-04-14, 2:46 am |
| In article <support-1E5AC1.01164614042005@news.verizon.net>,
flux <support@fluxsoft.com> wrote:
>In article <1113415288.189491@smirk>,
> _firstname_@lr_dot_los-gatos_dot_ca.us wrote:
>
>
>
>Your guess is wrong.
>
>
>There is no why because there probably isn't much difference.
And why, exactly, is that? Because you say so?
Doing sequential writes loads the mechanical and magnetic components of
a drive in an entirely different fashion than doing a random I/O load
including reads, writes, and long seeks. There is no rational reason to
assume that a continuous sequential-write load will cause the same type
of failures as a continuous random I/O load -- nor vice-versa.
As a trivial real-world example, in a continuous sequential-write load
like that of TiVo is is pretty much the case that every sector is
overwritten the same number of times. In a random I/O load through a
filesystem, blocks containing metadata will be rewritten many more times
than any _particular_ data block. The result of this is that if you have
a drive with cheap magnetic media in which individual sectors can only
be reliably overwritten a few tens or hundreds of thousands of times,
*which is in fact a common failure mode of some low-end desktop drives*,
the drive will fail much more quickly under the random I/O load than
under the TiVo load.
This is the sort of problem that those of us with actual experience with
large installations with many drives in servers with different workloads
have actually seen in practice; for example, I recently pulled every
Samsung SP040 out of an entire rack of servers because they had an
unacceptable failure rate even when paired in RAID 1 mirrors due to
exactly this issue. But, if I recall correctly, you're a college kid
posting from your dorm room; you have a lot of fancy theories and angry
talk but no actual experience with large installations in the field. I
see you've learned something about TiVo; did you also learn that DVR
manufacturers work closely with drive vendors to ensure that the drives
they ship have hardware and firmware carefully tuned for the particular
kind of continuous load that they require? One obvious example is that
many desktop drives do not depower the read electronics even when the
drive is not reading; this leads to increased wear, and decreased lifetime,
of the read head when the drive is placed in 100% duty cycle service; and
it is precisely the sort of thing that manufacturers tweak when tuning a
drive for a particular application (other such things involve using higher
quality magnetic media, as I mention above, changing load/unload behaviour,
and many more).
What _is_ nonsense is that all ATA (much less all SATA) drives are cheap
junk that can't be used in enterprise applications. Manufacturers make,
and warranty, a pretty good range of drives in both PATA and SATA now
that can be used in high duty cycle applications with a reasonable
degree of confidence, e.g. the Maxtor MAXline and WD Raptor and Raid
Edition drives. But these drives are most emphatically _not_ the same
in some ways as generic "desktop" ATA or SATA drives, and it is not
reasonable to say that all such drives will survive enterprise use,
though some may be built well enough to do so.
This also used to be true of SCSI drives; but the bottom dropped out of
the SCSI desktop drive market and left *only the high-end server
products* behind. The result is that though you can trust that _some_
particular SATA and ATA drives are designed, and tuned, for server use,
you can trust that just about all current SCSI drives are that way.
So there is a difference -- but it is not the difference most people
seem to think there is. Of course, it is also not the _lack_ of
difference you seem to insist there is; but since you seem to know
essentially nothing about anything, but enjoy talking very loud and
very often, that is not too surprising to me.
--
Thor Lancelot Simon tls@rek.tjls.com
"The inconsistency is startling, though admittedly, if consistency is to be
abandoned or transcended, there is no problem." - Noam Chomsky
| |
|
| In article <1113416697.696293@smirk>,
lr@idiom.com (Ralph Becker-Szendy) wrote:
> In article <fscp51hllq36ihajkrpv44dmmd7npbl0oh@4ax.com>,
> Curious George <cg@email.net> wrote:
>
> Furthermore, on real computers, power supplies and fans are all
> redundant, and can be hot-swapped. Look at the back of a good
But those things never seem to go down. Motherboards, RAM, CPUs do and
apparently with notably more frequency!
> This is where RAID comes in. But even that's pretty difficult. PATA
> was not designed for hot-swap; the fact that some disk enclosures and
> the 3ware cards can do it at all is a bit of a miracle. At least SATA
> is designed for hot swap. All SCSI and FC drives sold today are
> hot-swappable (that's why many of the SCSI drives are sold with the
> 80-pin connector that integrated data and power in one connector).
Wait, don't they need the SCA connector be hot-swappable? Most SCSI
drive do *not* have this connector. That would mean most of them are not
hot-swappable.
> In summary, the disks are and remain the least reliably component of a
> serious computer system.
In summary, disks, ATA or otherwise, seem to be really no less reliable
than any other component in the computer.
| |
| Thor Lancelot Simon 2005-04-14, 2:46 am |
| In article <support-7BD2A5.01491814042005@news.verizon.net>,
flux <support@fluxsoft.com> wrote:
>In article <1113416697.696293@smirk>,
> lr@idiom.com (Ralph Becker-Szendy) wrote:
>
>
>But those things never seem to go down. Motherboards, RAM, CPUs do and
>apparently with notably more frequency!
Apparently to _you_, perhaps. But all that does is confirm that you
plainly have very, very little experience -- at all, much less with
installations large enough to notice general trends.
Fans fail all the time. In fact, it's remarkably uncommon to see a | | |