On 07/05/2012 19:16, Bob Eager wrote:
On Mon, 07 May 2012 19:03:48 +0100, dennis@home wrote:

"John wrote in message
o.uk...
On 06/05/2012 22:09, dennis@home wrote:


"Bob wrote in message
...
On Sun, 06 May 2012 20:14:49 +0100, dennis@home wrote:

"Bob wrote in message
...

The point is, however, that modern disks have enough time, on a
powerfail, to retract the heads before the cushion fails.

Modern disks don't retract the heads like the old ones did. The old
ones
had ramps which the head carriers went up when the heads retracted,
this
removed them from the surface. Modern ones land on the disk surface,
they retract to a "safe" landing zone.

Of course, dennis. Just like the latest WD Caviar Black, which uses
ramp load.

That's one of the latest ones with shock protection isn't it? Not all
have that.

and yet a sentence earlier, you claimed that "Modern disks don't
retract the heads"... Does this not cause you cognitive dissonance?

No, why would it.
Anyway its easy to wriggle out of as you would have to prove no modern
disks land their heads to show what I said was actually untrue. However
you can read whatever you like into what I said even if you have to
stretch it to the limit.
I did actually forget that they have reintroduced the ramps on some
mobile disks as part of the shock protection.

The Caviar Black is not a mobile disk, though. And that has ramps.

Indeed...

The Green and Blue desktop 3.5" drives use the same arrangement

Nice clear view of them in here from about 1:27 on:

http://www.youtube.com/watch?v=uR1RO...feature=relmfu

Also good shot of the same arrangement on a Hitachi Deskstar:

http://www.youtube.com/watch?v=zGIaD9sNEWE

Contrast that with the on disk parking arrangement used by Seagate:

http://www.youtube.com/watch?v=3KwYO...eature=related


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
o.uk...
On 07/05/2012 19:03, dennis@home wrote:


"John Rumm" wrote in message
o.uk...
On 06/05/2012 22:09, dennis@home wrote:


"Bob Eager" wrote in message
...
On Sun, 06 May 2012 20:14:49 +0100, dennis@home wrote:

"Bob Eager" wrote in message
...

The point is, however, that modern disks have enough time, on a
powerfail, to retract the heads before the cushion fails.

Modern disks don't retract the heads like the old ones did. The old
ones
had ramps which the head carriers went up when the heads retracted,
this
removed them from the surface. Modern ones land on the disk surface,
they retract to a "safe" landing zone.

Of course, dennis. Just like the latest WD Caviar Black, which uses
ramp
load.

That's one of the latest ones with shock protection isn't it?
Not all have that.

and yet a sentence earlier, you claimed that "Modern disks don't
retract the heads"... Does this not cause you cognitive dissonance?

No, why would it.

Because the two are contradictory.

Anyway its easy to wriggle out of as you would have to prove no modern
disks land their heads to show what I said was actually untrue.

No, your logic is flawed...

Take it a step at a time, its quit easy. You said:

Modern disks don't retract the heads like the old ones did.

Yes, and if any modern disk doesn't then that logic is true.

You don't understand logic is what the evidence suggests.

"John Rumm" wrote in message
...

Contrast that with the on disk parking arrangement used by Seagate:

http://www.youtube.com/watch?v=3KwYO...eature=related

I don't see any loading ramps on that disk.
Thanks for showing that the logic of what I stated was true.

On Fri, 4 May 2012 11:02:08 +0000 (UTC),
(Andrew Gabriel) wrote:

Nowadays (and actually, even back then), I run OS that doesn't care

Which one(s)?

On 07/05/2012 19:47, dennis@home wrote:


"John Rumm" wrote in message
o.uk...
On 07/05/2012 19:03, dennis@home wrote:


"John Rumm" wrote in message
o.uk...
On 06/05/2012 22:09, dennis@home wrote:


"Bob Eager" wrote in message
...
On Sun, 06 May 2012 20:14:49 +0100, dennis@home wrote:

"Bob Eager" wrote in message
...

The point is, however, that modern disks have enough time, on a
powerfail, to retract the heads before the cushion fails.

Modern disks don't retract the heads like the old ones did. The old
ones
had ramps which the head carriers went up when the heads retracted,
this
removed them from the surface. Modern ones land on the disk surface,
they retract to a "safe" landing zone.

Of course, dennis. Just like the latest WD Caviar Black, which uses
ramp
load.

That's one of the latest ones with shock protection isn't it?
Not all have that.

and yet a sentence earlier, you claimed that "Modern disks don't
retract the heads"... Does this not cause you cognitive dissonance?

No, why would it.

Because the two are contradictory.

Anyway its easy to wriggle out of as you would have to prove no modern
disks land their heads to show what I said was actually untrue.

No, your logic is flawed...

Take it a step at a time, its quit easy. You said:

Modern disks don't retract the heads like the old ones did.

Yes, and if any modern disk doesn't then that logic is true.

You don't understand logic is what the evidence suggests.

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

"John Rumm" wrote in message
o.uk...


Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.

What mistake?
I said modern disks not every modern disk, I wouldn't claim that I knew how
every modern disk works unlike you appear to.
Its you that doesn't understand the logic so why not admit you have made a
mistake, not that anyone cares.
I admit that i could have included the word some to help you but I never
thought about helping you at the time.
I will try to think about the less fortunate in the future.

On Mon, 07 May 2012 23:42:32 +0100, dennis@home wrote:

"John Rumm" wrote in message
o.uk...


Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of
you.

What mistake?
I said modern disks not every modern disk, I wouldn't claim that I knew
how every modern disk works unlike you appear to. Its you that doesn't
understand the logic so why not admit you have made a mistake, not that
anyone cares.
I admit that i could have included the word some to help you but I never
thought about helping you at the time. I will try to think about the
less fortunate in the future.

Don't worry dennis, we understand that you deliberately use sloppy
English so that you always have space to wriggle.



--
Use the BIG mirror service in the UK:
http://www.mirrorservice.org

*lightning protection* - a w_tom conductor

On 07/05/2012 23:42, dennis@home wrote:


"John Rumm" wrote in message
o.uk...


Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.

What mistake?
I said modern disks not every modern disk, I wouldn't claim that I knew
how every modern disk works unlike you appear to.
Its you that doesn't understand the logic so why not admit you have made
a mistake, not that anyone cares.
I admit that i could have included the word some to help you but I never
thought about helping you at the time.
I will try to think about the less fortunate in the future.


Den you are just being deceitful...

Bob said

The point is, however, that modern disks have enough time, on a
powerfail, to retract the heads before the cushion fails.

To which you decided to have a pop with:

Modern disks don't retract the heads like the old ones did.

Now, that's a fairly unequivocal statement all by itself - no
qualifications with "many", or "some" or "those that I have seen" etc.

However when taken as a response to Bob's post, one can only conclude
you intended it to mean exactly how it reads - that *no* modern disk
retracts the heads.

Now if you maintain that you did not actually mean that after all, and
were not intending to speak about all disks, then we will gladly accept
your apology for sloppy and misleading wording instead, and move on.

So which is it to be?



--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

John Rumm wrote:

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.



He never has in the past


--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

"The Natural Philosopher" wrote in message
...
John Rumm wrote:

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.



He never has in the past

You haven't either with your claims about no OS
being able to handle a power failure when writing
and with head crashes due to power failure either.

"Rod Speed" wrote in message
...


"The Natural Philosopher" wrote in message
...
John Rumm wrote:

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.



He never has in the past

You haven't either with your claims about no OS
being able to handle a power failure when writing
and with head crashes due to power failure either.

True.

"Rod Speed" wrote in message
...


"The Natural Philosopher" wrote in message
...
John Rumm wrote:

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.



He never has in the past

You haven't either with your claims about no OS
being able to handle a power failure when writing
and with head crashes due to power failure either.

Ohhh look, the ozzie plonker. Full of **** but no substance.

Mike Tomlinson writes:

En el artÃ*culo , The Natural Philosopher
escribió:

No OS can cope with a power loss in the middle of a disk write operation.

Look up journalling filesystems. examples: ext3 and NTFS.

So long as you have barriers enabled and/or the disk's write cache
disabled. Up to 64MB of data that the O/S thinks has been flushed but
the disk hasn't yet written isn't good.

--
Alan J. Wylie http://www.wylie.me.uk/

"John Rumm" wrote in message
...

Now if you maintain that you did not actually mean that after all, and
were not intending to speak about all disks, then we will gladly accept
your apology for sloppy and misleading wording instead, and move on.

So which is it to be?

Like I said you got the logic wrong, now man up and admit it.
I didn't say what you assumed.
You can argue that I could have been clearer if you like but it was you that
mistakenly read it in the first place.

It was also correct that some modern disks don't unload the heads from the
platters as i have already clarified.

Anyway you can stop arguing about your misunderstanding now as I don't
really care why you got the logic wrong.

"The Natural Philosopher" wrote in message
...
John Rumm wrote:

Man up den, and stop behaving like a school kid. Just admit you made a
mistake, there is no shame in that, and we will all think better of you.



He never has in the past

You aren't very bright are you? I already have said it wasn't perfectly
clear.
however that doesn't stop the logic being correct and john being wrong even
if he refuses to admit it.

On Mon, 07 May 2012 11:33:42 +0100, Mike Tomlinson
wrote:

En el artículo m,
brass monkey escribió:

reams of your puerile **** any 2 year old could leave for dead flushed
where it belongs

LOL, hilarious
What a plonker.

When Woddles is losing an argument, he blows a gasket, morphs in a
failed attempt to evade killfiles and posts boilerplate insults.
Plus ca change.

Rod Speed FAQ: http://tinyurl.com/883xp7v


http://www.sensationbot.com/jschat.php?db=rodspeed

excellent!

Jim K

"Jim K" wrote in message
news On Mon, 07 May 2012 11:33:42 +0100, Mike Tomlinson
wrote:

En el artículo m,
brass monkey escribió:

reams of your puerile **** any 2 year old could leave for dead flushed
where it belongs

LOL, hilarious
What a plonker.

When Woddles is losing an argument, he blows a gasket, morphs in a
failed attempt to evade killfiles and posts boilerplate insults.
Plus ca change.

Rod Speed FAQ: http://tinyurl.com/883xp7v


http://www.sensationbot.com/jschat.php?db=rodspeed

excellent!

Classic

Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then obviously not.

If you mean "the file system will not be corrupted, and the write will either appear to have completed, or not started" - then I believe there are various operating systems that can do exactly that.


If power fails during a write, data is always corrupted: the question is
by how much, whether its recoverable, and what performance hit you take
on trying to avoid it.


e.g writing to two mirror disks, one delayed and only written to when
the other is NOT writing should give you at least one file system that
is not corrupted. If somewhat stale.



Of course, there is then the question of whether the applications running on top of the OS offer the same guarantees.


--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then
obviously not.

If you mean "the file system will not be corrupted, and the write will
either appear to have completed, or not started" - then I believe
there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been told is
written will be even though it's not necessarily been put onto the
magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?


--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

On 05/05/2012 20:05, dennis@home wrote:

Hardware bombs are nothing to do with the OS surviving a power cut.
Indeed, but lots of kit will run happily for years *until* it gets
turned off.

On 08/05/2012 12:06, brass monkey wrote:
"Jim wrote in message
news On Mon, 07 May 2012 11:33:42 +0100, Mike
wrote:

En el raweb.com,
brass escribió:

reams of your puerile **** any 2 year old could leave for dead flushed
where it belongs

LOL, hilarious
What a plonker.

When Woddles is losing an argument, he blows a gasket, morphs in a
failed attempt to evade killfiles and posts boilerplate insults.
Plus ca change.

Rod Speed FAQ: http://tinyurl.com/883xp7v


http://www.sensationbot.com/jschat.php?db=rodspeed

excellent!

Classic

You can actually get more sense out of the bot than you do the real one!


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

John Rumm wrote:
On 08/05/2012 12:06, brass monkey wrote:
"Jim wrote in message
news On Mon, 07 May 2012 11:33:42 +0100, Mike
wrote:

En el traweb.com,
brass escribió:

reams of your puerile **** any 2 year old could leave for dead
flushed
where it belongs

LOL, hilarious
What a plonker.

When Woddles is losing an argument, he blows a gasket, morphs in a
failed attempt to evade killfiles and posts boilerplate insults.
Plus ca change.

Rod Speed FAQ: http://tinyurl.com/883xp7v


http://www.sensationbot.com/jschat.php?db=rodspeed

excellent!

Classic

You can actually get more sense out of the bot than you do the real one!


It is uncanningly accurate.

You: do you think that Freud was correct about youir sexual hangups?
Rod Speed: It might well be if it was more than just your pathetic
little fantasy.
You: Ah, so dribbling in your gargle glass is not a sign of senility?
Rod Speed: I wont be bothering to reply in this thread again, too
boring, if you are too stupid to understand, thats entirely your problem.
You: Are you able to actually respond intelligently?
Rod Speed: No one actually gives a frf what you'd rather boy. You made
the claims, you get to substantiate those claims. Thats how it works.



--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

On 08/05/2012 10:08, dennis@home wrote:


"John Rumm" wrote in message
...

Now if you maintain that you did not actually mean that after all, and
were not intending to speak about all disks, then we will gladly
accept your apology for sloppy and misleading wording instead, and
move on.

So which is it to be?

Like I said you got the logic wrong, now man up and admit it.

The logic was fine - the problem seems to rest either with your
integrity, or your English language skills.

(Generally we assume the former, because we know that with anything you
say, you may attempt to revise its meaning a-posteriori as it best suits
your argument)

I didn't say what you assumed.

What you *said* was very clear:

"Modern disks don't retract the heads like the old ones did."

However you now appear to be suggesting that we should have interpreted
that that as "Some modern disks do, and some don't". Alas that does not
make sense with the "like the old ones did." qualification you added.

You can argue that I could have been clearer if you like

One could indeed. Alternatively we could go with the more likely
explanation that you simply posted something incorrect, and won't admit
it even at the cost of now appearing absurd.

but it was you
that mistakenly read it in the first place.

I read what you wrote - not what you now claim you meant.

Personally I think you simply made an incorrect claim without checking
your facts, and are now attempting to wiggle out as normal.

It was also correct that some modern disks don't unload the heads from
the platters

Don't think anyone was arguing otherwise.

as i have already clarified.

ITYM as has now been clarified for you.

Anyway you can stop arguing about your misunderstanding now as I don't
really care why you got the logic wrong.

You should care for two reasons. Firstly, if there was a
misunderstanding, then it resulted from your sloppy wording. Secondly if
your interpretation of the statement "Modern disks don't retract the
heads like the old ones did." is actually that "Modern disks can retract
the heads like the old ones did.", then you have serious comprehension
issues.

--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On 08/05/2012 15:18, The Natural Philosopher wrote:
Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then
obviously not.

If you mean "the file system will not be corrupted, and the write
will either appear to have completed, or not started" - then I
believe there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been told
is written will be even though it's not necessarily been put onto the
magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?

The transaction is not marked as complete in the NVRAM, and hence is
never committed... potentially resulting in data loss, but not low level
file system corruption.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On 08/05/2012 18:29, John Rumm wrote:
On 08/05/2012 15:18, The Natural Philosopher wrote:
Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then
obviously not.

If you mean "the file system will not be corrupted, and the write
will either appear to have completed, or not started" - then I
believe there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been told
is written will be even though it's not necessarily been put onto the
magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?

The transaction is not marked as complete in the NVRAM, and hence is
never committed... potentially resulting in data loss, but not low level
file system corruption.

And the client hasn't been told that the data is written, so it can do
whatever is appropriate. Of course it's likely the client will be in the
same data centre and be too busy falling over itself to do anything else :-)

"Chris Bartram" wrote in message
...
On 05/05/2012 20:05, dennis@home wrote:

Hardware bombs are nothing to do with the OS surviving a power cut.
Indeed, but lots of kit will run happily for years *until* it gets turned
off.

I know.
I once installed a xenix OS running on a 286 for a project.
We used it for about 2.5 years and then finished the project.
We forgot to turn it off and it was found ~12 years later sitting in the
bottom of a rack the IT department when they wanted to move the rack.
it was still up and running without a reboot.

Oscar wrote:


"The Natural Philosopher" wrote in message
...
Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then
obviously not.

If you mean "the file system will not be corrupted, and the write
will either appear to have completed, or not started" - then I
believe there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been told
is written will be even though it's not necessarily been put onto the
magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?

The write to NVRAM happens so quickly that there is plenty of time
to do that before the supercap powering the NVRAM goes.



I see you have very little clue about computer hardware

--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.



--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

On 08/05/2012 02:55, brass monkey wrote:
Ohhh look, the ozzie plonker. Full of **** but no substance.


I bet you wouldn't say that about a Pakistani.

Andy

"The Natural Philosopher" wrote in message
...
Oscar wrote:


"The Natural Philosopher" wrote in message
...
Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher
wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" - then
obviously not.

If you mean "the file system will not be corrupted, and the write will
either appear to have completed, or not started" - then I believe
there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been told is
written will be even though it's not necessarily been put onto the
magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?

The write to NVRAM happens so quickly that there is plenty of time
to do that before the supercap powering the NVRAM goes.



I see you have very little clue about computer hardware

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.

--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.



--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.

"Oscar" wrote in message
...


"The Natural Philosopher" wrote in message
...

8

so what happens when you are writing to NVRAM and the power goes down?

The write to NVRAM happens so quickly that there is plenty of time
to do that before the supercap powering the NVRAM goes.

Does the supercap keep the CPU + memory, etc. going for the time it takes to
write a full allocation unit + allocation table updates to the NVRAM? Or are
we just talking about updating the intent logs?
Its not a byte or two.

Oscar wrote:

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption
on power failure and how. Unless that post got lost in the noise, that
is. All the ways round it that I've seen mentioned involve NAS and
caches of some sort.

--
Tciao for Now!

John.

"John Williamson" wrote in message
...
Oscar wrote:

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption on
power failure and how. Unless that post got lost in the noise, that is.
All the ways round it that I've seen mentioned involve NAS and caches of
some sort.

Disc CORRUPTION is not the same thing as ensuring that all data
gets written to the platter.

--
Tciao for Now!

John.

On 08/05/2012 21:09, Andy Champ wrote:
On 08/05/2012 02:55, brass monkey wrote:
Ohhh look, the ozzie plonker. Full of **** but no substance.


I bet you wouldn't say that about a Pakistani.

I am sure he would if that is where wodney came from...

(the salient point being the plonker bit rather than the location -
although it does distinguish from one of our home grown ones)


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

Oscar wrote:


"John Williamson" wrote in message
...
Oscar wrote:

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption
on power failure and how. Unless that post got lost in the noise, that
is. All the ways round it that I've seen mentioned involve NAS and
caches of some sort.

Disc CORRUPTION is not the same thing as ensuring that all data
gets written to the platter.

It is if the power failure happens while you're writing the FAT.


--
Tciao for Now!

John.

On Tue, 08 May 2012 22:33:57 +0100, John Williamson wrote:

Oscar wrote:


"John Williamson" wrote in message
...
Oscar wrote:

Everyone has shown that you have none on that question of OSs being
quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption
on power failure and how. Unless that post got lost in the noise, that
is. All the ways round it that I've seen mentioned involve NAS and
caches of some sort.

Disc CORRUPTION is not the same thing as ensuring that all data gets
written to the platter.

It is if the power failure happens while you're writing the FAT.

But there are usually two...one will survive.

--
Use the BIG mirror service in the UK:
http://www.mirrorservice.org

*lightning protection* - a w_tom conductor

Bob Eager wrote:
On Tue, 08 May 2012 22:33:57 +0100, John Williamson wrote:

Oscar wrote:

"John Williamson" wrote in message
...
Oscar wrote:
Everyone has shown that you have none on that question of OSs being
quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption
on power failure and how. Unless that post got lost in the noise, that
is. All the ways round it that I've seen mentioned involve NAS and
caches of some sort.
Disc CORRUPTION is not the same thing as ensuring that all data gets
written to the platter.

It is if the power failure happens while you're writing the FAT.

But there are usually two...one will survive.

Which is why the spec says two FATs. But you still need to find out
which FAT matches the HD contents, and until you do that, the HD is
unusable. You seem to be saying that as long as the data is recoverable
by one means or another, the HD isn't corrupt in your opinion.

--
Tciao for Now!

John.

On Tue, 08 May 2012 22:43:30 +0100, John Williamson wrote:

Bob Eager wrote:
On Tue, 08 May 2012 22:33:57 +0100, John Williamson wrote:

Oscar wrote:

"John Williamson" wrote in message
...
Oscar wrote:
Everyone has shown that you have none on that question of OSs being
quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD
corruption on power failure and how. Unless that post got lost in
the noise, that is. All the ways round it that I've seen mentioned
involve NAS and caches of some sort.
Disc CORRUPTION is not the same thing as ensuring that all data gets
written to the platter.

It is if the power failure happens while you're writing the FAT.

But there are usually two...one will survive.

Which is why the spec says two FATs. But you still need to find out
which FAT matches the HD contents, and until you do that, the HD is
unusable. You seem to be saying that as long as the data is recoverable
by one means or another, the HD isn't corrupt in your opinion.

Perfectly possible with a timestamp on each FAT..although not done very
often!

--
Use the BIG mirror service in the UK:
http://www.mirrorservice.org

*lightning protection* - a w_tom conductor

"John Williamson" wrote in message
...
Oscar wrote:


"John Williamson" wrote in message
...
Oscar wrote:

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD corruption
on power failure and how. Unless that post got lost in the noise, that
is. All the ways round it that I've seen mentioned involve NAS and
caches of some sort.

Disc CORRUPTION is not the same thing as ensuring that all data
gets written to the platter.

It is if the power failure happens while you're writing the FAT.

The OSs that have file systems that avoid corruption on a power failure
do not use a simple FAT.


--
Tciao for Now!

John.

On Wed, 09 May 2012 08:05:04 +1000, Oscar wrote:

"John Williamson" wrote in message
...
Oscar wrote:


"John Williamson" wrote in message
...
Oscar wrote:

Everyone has shown that you have none on that question of OSs being
quite capable of avoiding disc corruption on power failure.

I notice that you still haven't told us which OS avoided HD
corruption on power failure and how. Unless that post got lost in the
noise, that is. All the ways round it that I've seen mentioned
involve NAS and caches of some sort.

Disc CORRUPTION is not the same thing as ensuring that all data gets
written to the platter.

It is if the power failure happens while you're writing the FAT.

The OSs that have file systems that avoid corruption on a power failure
do not use a simple FAT.

Indeed. Even back in the 1979s, there were file systems (at least for
experimental purposes) that were just careful about the order of metadata
updates.

--
Use the BIG mirror service in the UK:
http://www.mirrorservice.org

*lightning protection* - a w_tom conductor

Oscar wrote:


"The Natural Philosopher" wrote in message
...
Oscar wrote:


"The Natural Philosopher" wrote in message
...
Clive George wrote:
On 08/05/2012 13:50, Martin Bonner wrote:
On Saturday, 5 May 2012 00:25:53 UTC+1, The Natural Philosopher
wrote:
No OS can cope with a power loss in the middle of a
disk write operation.
Rather depends what you mean by "cope with".

If you mean "the data you were trying to write gets written" -
then obviously not.

If you mean "the file system will not be corrupted, and the write
will either appear to have completed, or not started" - then I
believe there are various operating systems that can do exactly that.

Of course, there is then the question of whether the applications
running on top of the OS offer the same guarantees.

Proper SAN/NAS places the write into NVRAM or battery-backed RAM or
similar before it goes to disk, so anything the client has been
told is written will be even though it's not necessarily been put
onto the magnetic bits yet.

so what happens when you are writing to NVRAM and the power goes down?

The write to NVRAM happens so quickly that there is plenty of time
to do that before the supercap powering the NVRAM goes.



I see you have very little clue about computer hardware

Everyone has shown that you have none on that question of OSs
being quite capable of avoiding disc corruption on power failure.


That is because there is no way to avoid it without the hardware
actually being connected to a power supply that pre-warns of impending
voltage loss and actually cleans up fast before losing power.


--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.



--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.



--
To people who know nothing, anything is possible.
To people who know too much, it is a sad fact
that they know how little is really possible -
and how hard it is to achieve it.