Where is a tmpfs physically placed? In memory RAM or on hard disc?

[Peter Hanke]

Assume I define in file /etc/fstab a new line for an additional tmpfs filesystem with mount point /tmp
(and mount it either manually of through reboot).

Then I copy some file stuff onto this filesystem dir /tmp

Where are these files physically placed? Does /tmp only exist in memory RAM or inside an existing partition (which ?) on hard disc?

Peter

[Tauno Voipio]

Did you attempt to google for 'linux tmpfs'?

The first few responses provided the information:

tmpfs is a common name for a temporary file storage facility on many
Unix-like operating systems. It is intended to appear as a mounted file
system, but stored in volatile memory instead of a persistent storage
device. A similar construction is a RAM disk, which appears as a virtual
disk drive and hosts a disk file system.

--

Tauno Voipio

[David Brown]

Your files are typically only in memory. The tmpfs storage is swapable,
so the data can physically migrate into swap partitions if that suits
memory requirements better (but the files will never be stored as
"files" in a filesystem in the swap partitions - bits of them will just
get paged out here and there).

[Casper H.S. Dik]

Anywhere. They might be moved to disk when the files aren't being used
and the kernel pages the memory. They are not stored in a partition except
in swap partitions/files.

Casper

[Aragorn]

On Friday 16 November 2012 07:13, Peter Hanke conveyed the following to
comp.os.linux.misc...

tmpfs is a pageable RAM filesystem, so the files will be placed in RAM
until the upper limit of the tmpfs is reached - this is half of your RAM
by default if you do not specify another limit - in which case the
contents of the tmpfs will then be paged out to the swap partition, just
as with the other contents of userspace memory.

Do however keep in mind that if there are any files in /tmp before you
mount the tmpfs on it, then these files will still persist, but will be
unreachable as long as /tmp has another filesystem mounted to it,
whether that's a tmpfs or a physical partition on your disk.

By consequence, the best modus operandi is to drop down to single-user
maintenance mode - alias "runlevel 1" - and then make sure that /tmp is
empty. Then edit your /etc/fstab, save it, mount a tmpfs over /tmp, and
return to your normal runlevel.

--
= Aragorn =
(registered GNU/Linux user #223157)

[David Brown]

On 16/11/2012 14:12, Aragorn wrote:
> On Friday 16 November 2012 07:13, Peter Hanke conveyed the following to
> comp.os.linux.misc...
>
>> Assume I define in file /etc/fstab a new line for an additional tmpfs
>> filesystem with mount point /tmp (and mount it either manually of
>> through reboot).
>>
>> Then I copy some file stuff onto this filesystem dir /tmp
>>
>> Where are these files physically placed? Does /tmp only exist in
>> memory RAM or inside an existing partition (which ?) on hard disc?
>
> tmpfs is a pageable RAM filesystem, so the files will be placed in RAM
> until the upper limit of the tmpfs is reached - this is half of your RAM
> by default if you do not specify another limit - in which case the
> contents of the tmpfs will then be paged out to the swap partition, just
> as with the other contents of userspace memory.

Not quite true.

tmpfs will take as much or as little memory for all the files as it
needs, up to the limit you set when mounting it. This limit is half
your physical ram size by default, but you can pick a different value
when mounting, and you can change it while it is in use (using a "mount
-o remount" command).

Whether or not parts of the tmpfs are swapped out to disk is independent
of the size limit of the tmpfs and the amount that is actually in use.
Quite simply, the storage space is just normal allocated swappable ram -
if pages in the tmpfs are left untouched, they can get swapped out if
the OS thinks it has better use for the memory.

>
> Do however keep in mind that if there are any files in /tmp before you
> mount the tmpfs on it, then these files will still persist, but will be
> unreachable as long as /tmp has another filesystem mounted to it,
> whether that's a tmpfs or a physical partition on your disk.
>
> By consequence, the best modus operandi is to drop down to single-user
> maintenance mode - alias "runlevel 1" - and then make sure that /tmp is
> empty. Then edit your /etc/fstab, save it, mount a tmpfs over /tmp, and
> return to your normal runlevel.
>

It is far easier for non-experts to just edit /etc/fstab, then reboot.

[unruh]

In memory Ram. It disappears if the computer is switched off.

>
> Peter
>

[Balwinder S Dheeman]

See also: http://www.kernel.org/doc/Documentation/filesystems/tmpfs.txt

--
Balwinder S "bdheeman" Dheeman
(http://werc.homelinux.net/contact/)

[Pascal Hambourg]

Hello,

unruh a ï¿œcrit :

> On 2012-11-16, Peter Hanke <pete...@andres.net> wrote:
>> Assume I define in file /etc/fstab a new line for an additional tmpfs filesystem with mount point /tmp
>> (and mount it either manually of through reboot).
>>
>> Then I copy some file stuff onto this filesystem dir /tmp
>>
>> Where are these files physically placed? Does /tmp only exist in memory RAM or inside an existing partition (which ?) on hard disc?
>
> In memory Ram.

No, in virtual memory, i.e. either in system RAM or swap space.

> It disappears if the computer is switched off.

Swap space contents does not disappear when the computer is switched
off. Even RAM contents does not disappear immediately.

[Richard Kettlewell]

Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
> unruh a écrit :

>> It disappears if the computer is switched off.
>
> Swap space contents does not disappear when the computer is switched
> off. Even RAM contents does not disappear immediately.

The sense in which it disappears is that you’re not going to get the
contents back when you restore power.

--
http://www.greenend.org.uk/rjk/

[The Natural Philosopher]

..unless you suspend before switching off, when IIRC the whole RAM area
gets 'swapped out' somewhere.

I love the reboot on the windows VMM..less than 10 seconds from cold to
'where I last left windows two weeks ago when I used it'

--
Ineptocracy

(in-ep-toc’-ra-cy) – a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

[Pascal Hambourg]

Richard Kettlewell a ï¿œcrit :
> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>> unruh a ï¿œcrit :

>
>>> It disappears if the computer is switched off.
>> Swap space contents does not disappear when the computer is switched
>> off. Even RAM contents does not disappear immediately.
>
> The sense in which it disappears is that you're not going to get the
> contents back when you restore power.

I would not rely on such assomption if I needed to make sure that the
data actually disappears.
E.g. see <http://en.wikipedia.org/wiki/Data_remanence#Data_in_RAM>
Not to mention the data that were swapped out.

[Aragorn]

On Saturday 17 November 2012 23:20, The Natural Philosopher conveyed the
following to comp.os.linux.misc...

> I love the reboot on the windows VMM..less than 10 seconds from cold
> to 'where I last left windows two weeks ago when I used it'

What I love the most about Microsoft Windows is the fact that I don't
have it installed anywhere. :-]

[unruh]

On 2012-11-17, The Natural Philosopher <t...@invalid.invalid> wrote:
> On 17/11/12 21:51, Richard Kettlewell wrote:
>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:

>>> unruh a ??crit :

>>
>>>> It disappears if the computer is switched off.
>>>
>>> Swap space contents does not disappear when the computer is switched
>>> off. Even RAM contents does not disappear immediately.
>>

>> The sense in which it disappears is that you???re not going to get the

>> contents back when you restore power.
>>
> ..unless you suspend before switching off, when IIRC the whole RAM area
> gets 'swapped out' somewhere.

I think you mean hibernate, not suspend.

>
> I love the reboot on the windows VMM..less than 10 seconds from cold to
> 'where I last left windows two weeks ago when I used it'

You can hibernate Linux as well. and yes, about 10 sec from cold to
where I left Linux at 2 hrs ago when I last used it.

[unruh]

On 2012-11-17, Pascal Hambourg <boite-...@plouf.fr.eu.org> wrote:
> Richard Kettlewell a ?crit :
>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>>> unruh a ?crit :

>>
>>>> It disappears if the computer is switched off.
>>> Swap space contents does not disappear when the computer is switched
>>> off. Even RAM contents does not disappear immediately.
>>
>> The sense in which it disappears is that you're not going to get the
>> contents back when you restore power.
>
> I would not rely on such assomption if I needed to make sure that the
> data actually disappears.
> E.g. see <http://en.wikipedia.org/wiki/Data_remanence#Data_in_RAM>
> Not to mention the data that were swapped out.

Uh, there is nothing in his post which suggests that he is worried about
the security implications of using tempfs. If he is, the best way is
when you switch off your computer your first run a road roller over it,
then place it in a blast furnace at 2000C.

[Aragorn]

On Sunday 18 November 2012 00:16, unruh conveyed the following to
comp.os.linux.misc...

I would suggest adding a third step to those two, and before your first
one, namely dipping it in liquid nitrogen first. :-D

[Pascal Hambourg]

unruh a ï¿œcrit :

> On 2012-11-17, Pascal Hambourg <boite-...@plouf.fr.eu.org> wrote:
>> Richard Kettlewell a ?crit :
>>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>>>> unruh a ?crit :
>>>>> It disappears if the computer is switched off.
>>>> Swap space contents does not disappear when the computer is switched
>>>> off. Even RAM contents does not disappear immediately.
>>> The sense in which it disappears is that you're not going to get the
>>> contents back when you restore power.
>> I would not rely on such assomption if I needed to make sure that the
>> data actually disappears.
>> E.g. see <http://en.wikipedia.org/wiki/Data_remanence#Data_in_RAM>
>> Not to mention the data that were swapped out.
>
> Uh, there is nothing in his post which suggests that he is worried about
> the security implications of using tempfs.

There is nothing either suggesting otherwise. Better be warned.
This is a kind of "fuzzy logic" where you always lose :
- if you expect that the data is still present, it may have disappeared
- if you expect that the data has disappeared, it may still be present.

> If he is, the best way is
> when you switch off your computer your first run a road roller over it,
> then place it in a blast furnace at 2000C.

What about erasing the system memory before switching off instead ?
Sounds cheaper.

[unruh]

On 2012-11-17, Pascal Hambourg <boite-...@plouf.fr.eu.org> wrote:

> unruh a ?crit :

You wanted him to be sure.

[The Natural Philosopher]

On 17/11/12 23:02, Aragorn wrote:
> On Saturday 17 November 2012 23:20, The Natural Philosopher conveyed the
> following to comp.os.linux.misc...
>
>> I love the reboot on the windows VMM..less than 10 seconds from cold
>> to 'where I last left windows two weeks ago when I used it'
>
> What I love the most about Microsoft Windows is the fact that I don't
> have it installed anywhere. :-]
>

I wish I could do without RhinoCAD..

[The Natural Philosopher]

Golly. you go as long as tow hours without using it?
:-)

[Pascal Hambourg]

unruh a écrit :

> On 2012-11-17, Pascal Hambourg <boite-...@plouf.fr.eu.org> wrote:
>> unruh a ?crit :
>>

>>> If he is, the best way is
>>> when you switch off your computer your first run a road roller over it,
>>> then place it in a blast furnace at 2000C.
>> What about erasing the system memory before switching off instead ?
>> Sounds cheaper.
>
> You wanted him to be sure.

No, I just wanted him (or any reader) to be aware, in case it is an issue.

[Richard Kettlewell]

Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
> Richard Kettlewell a écrit :
>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>>> unruh a écrit :

>>>> It disappears if the computer is switched off.
>>> Swap space contents does not disappear when the computer is switched
>>> off. Even RAM contents does not disappear immediately.
>>
>> The sense in which it disappears is that you're not going to get the
>> contents back when you restore power.
>
> I would not rely on such assomption if I needed to make sure that the
> data actually disappears.
> E.g. see <http://en.wikipedia.org/wiki/Data_remanence#Data_in_RAM>
> Not to mention the data that were swapped out.

I don’t recall anyone saying they did rely on that assumption. In any
case if your confidential data is sufficiently valuable that this kind
of attack is plausible then you need to do more than just be cautious
about its persistence in RAM after power off. Most people have more
realistic threats to worry about.

--
http://www.greenend.org.uk/rjk/

[Pascal Hambourg]

Richard Kettlewell a écrit :
> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>> Richard Kettlewell a écrit :
>>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>>>> unruh a écrit :
>
>>>>> It disappears if the computer is switched off.
>>>> Swap space contents does not disappear when the computer is switched
>>>> off. Even RAM contents does not disappear immediately.
>>> The sense in which it disappears is that you're not going to get the
>>> contents back when you restore power.
>> I would not rely on such assomption if I needed to make sure that the
>> data actually disappears.
>> E.g. see <http://en.wikipedia.org/wiki/Data_remanence#Data_in_RAM>
>> Not to mention the data that were swapped out.
>
> I don't recall anyone saying they did rely on that assumption.

I don't recall either anyone saying they did not. So in doubt...

> In any
> case if your confidential data is sufficiently valuable that this kind
> of attack is plausible then you need to do more than just be cautious
> about its persistence in RAM after power off. Most people have more
> realistic threats to worry about.

Nice try. But you were the one who wrote "The sense in which it

disappears is that you're not going to get the contents back when you

restore power", and that was proved wrong in some occasions. That was my
one and only point.

[The Natural Philosopher]

[The Natural Philosopher]

On 18/11/12 14:27, Pascal Hambourg wrote:

You wont get the RAM contents back when you restore power except by
using some advanced and careful technology probably involving removing
the RAM and putting it in another machine altogether which is designed
NOT to set it all to zeroes when it boots, or overwrite it with the
operating system.

"With additional hardware and software, an Apple LISA could send a
rocket to the moon"

[The Natural Philosopher]

Sorry. Fat finger causes null post.

[unruh]

On 2012-11-18, The Natural Philosopher <t...@invalid.invalid> wrote:
> On 17/11/12 23:14, unruh wrote:
>> On 2012-11-17, The Natural Philosopher <t...@invalid.invalid> wrote:
>>> On 17/11/12 21:51, Richard Kettlewell wrote:
>>>> Pascal Hambourg <boite-...@plouf.fr.eu.org> writes:
>>>>> unruh a ??crit :
>>>>
>>>>>> It disappears if the computer is switched off.
>>>>>
>>>>> Swap space contents does not disappear when the computer is switched
>>>>> off. Even RAM contents does not disappear immediately.
>>>>
>>>> The sense in which it disappears is that you???re not going to get the
>>>> contents back when you restore power.
>>>>
>>> ..unless you suspend before switching off, when IIRC the whole RAM area
>>> gets 'swapped out' somewhere.
>>
>> I think you mean hibernate, not suspend.
>>>
>>> I love the reboot on the windows VMM..less than 10 seconds from cold to
>>> 'where I last left windows two weeks ago when I used it'
>>
>> You can hibernate Linux as well. and yes, about 10 sec from cold to
>> where I left Linux at 2 hrs ago when I last used it.
>>
>>
> Golly. you go as long as tow hours without using it?

And even two hours. Yes, that is how long the bus ride from work is. I
have even been known to go for a whole day.
One problem with hibernation on my laptop is that the computer comes out
of hibernation if the lid is opened ( a small way) I have had the system
come out of hibernation in my case I guess when it got jolted. Now I
briefly remove the battery as well after hibernation to make sure it is
off. (Running a laptop inside a well padded case does not do it any
good.)


>:-)
>

[Aragorn]

On Sunday 18 November 2012 15:31, The Natural Philosopher conveyed the
following to comp.os.linux.misc...

> "With additional hardware and software, an Apple LISA could send a
> rocket to the moon"

Didn't the first flight to the moon make use of two Commodore C64
machines?

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> The Natural Philosopher conveyed the following to comp.os.linux.misc...

>> "With additional hardware and software, an Apple LISA could send a
>> rocket to the moon"
>
> Didn't the first flight to the moon make use of two Commodore C64
> machines?

Not unless NASA had a time machine...

--
http://www.greenend.org.uk/rjk/

[Aragorn]

On Sunday 18 November 2012 20:43, Richard Kettlewell conveyed the

following to comp.os.linux.misc...

> Aragorn <str...@telenet.be.invalid> writes:
>
>> The Natural Philosopher conveyed the following to
>> comp.os.linux.misc...
>
>>> "With additional hardware and software, an Apple LISA could send a
>>> rocket to the moon"
>>
>> Didn't the first flight to the moon make use of two Commodore C64
>> machines?
>
> Not unless NASA had a time machine...

You are correct. Wikipedia states that the C64 was introduced in 1982.

I do however maintain that NASA does, indeed, have a time machine. I
will however not insist that it would bear any physical resemblance to a
DeLorean DMC-12.

[Richard Kettlewell]

The Natural Philosopher <t...@invalid.invalid> writes:
> "With additional hardware and software, an Apple LISA could send a
> rocket to the moon"

“Do not forget that, with China, Albania has a quarter of the world’s
population”. (Allegedly Enver Hoxha, though I have my doubts.)

--
http://www.greenend.org.uk/rjk/

[Michael Black]

On Sun, 18 Nov 2012, Aragorn wrote:

> On Sunday 18 November 2012 15:31, The Natural Philosopher conveyed the
> following to comp.os.linux.misc...
>
>> "With additional hardware and software, an Apple LISA could send a
>> rocket to the moon"
>
> Didn't the first flight to the moon make use of two Commodore C64
> machines?
>

That's an outright lie, to make NASA look bad.

What really is the case was that the original flight to the moon used
computers that were about equivalent to the C64.

NASA had massive mainframes on earth, and likely minicomputers too, and
none would amount to much compared with what we have in the home right
now, likely wasn't that much better than the C64 that came out 13 years
after the first moon landing.

I can't even remember if they had a computer on board Apollo 11. If it
was, it would be very minor, likely following a sequence of instructions
rather than actually computer.

Later flights, Skylab and maybe the joint US-USSR mission in the
seventies, they at least likely had the use of pocket calculators that
came along in the interim.

The shuttle's computer came later, but not that much later since the
design was frozen long before it made it into space. I seem to recal
something like core memory, and it wasn't that great a computer. But once
in place, it took tremendous effort to change, likely was risky sine
someone might misread something and create software that sent them off on
a tangent or something. And it usually does take forever to modify
"airplanes", a lot peripheral worry like does it interfere with other
things, and needing to fit a preassigned space.

But it wasn't that many years after the shuttle first went up that they
took laptops, I guess it was the Grid, which certainly was better than the
on board computer, even if it was quite limited compared to what I'm using
right now.

Michael

[Aragorn]

On Sunday 18 November 2012 22:00, Michael Black conveyed the following

to comp.os.linux.misc...

> On Sun, 18 Nov 2012, Aragorn wrote:
>
>> On Sunday 18 November 2012 15:31, The Natural Philosopher conveyed
>> the following to comp.os.linux.misc...
>>
>>> "With additional hardware and software, an Apple LISA could send a
>>> rocket to the moon"
>>
>> Didn't the first flight to the moon make use of two Commodore C64
>> machines?
>>
> That's an outright lie, to make NASA look bad.

Or to make the Commodore 64 look very good. :-D

> What really is the case was that the original flight to the moon used
> computers that were about equivalent to the C64.
>
> NASA had massive mainframes on earth, and likely minicomputers too,
> and none would amount to much compared with what we have in the home
> right now, likely wasn't that much better than the C64 that came out
> 13 years after the first moon landing.
>
> I can't even remember if they had a computer on board Apollo 11. If
> it was, it would be very minor, likely following a sequence of
> instructions rather than actually computer.

Something to that effect, which is why I believed the story about the
Commodores. Uhh, the computers, not the band. :p

> Later flights, Skylab and maybe the joint US-USSR mission in the
> seventies, they at least likely had the use of pocket calculators that
> came along in the interim.
>
> The shuttle's computer came later, but not that much later since the
> design was frozen long before it made it into space. I seem to recal
> something like core memory, and it wasn't that great a computer. But
> once in place, it took tremendous effort to change, likely was risky
> sine someone might misread something and create software that sent
> them off on a tangent or something. And it usually does take forever
> to modify "airplanes", a lot peripheral worry like does it interfere
> with other things, and needing to fit a preassigned space.
>
> But it wasn't that many years after the shuttle first went up that
> they took laptops, I guess it was the Grid, which certainly was better
> than the on board computer, even if it was quite limited compared to
> what I'm using right now.

Oh, that's not surprising. Any recent cellphone - it doesn't even have
to be a smartphone - has more computing power than my dad's pre-owned
hard-disk-less [*] XT clone in 1990. ;-)


[*] My brother and I did buy him a hard disk for that machine later on,
but to my dad, that machine was just a toy anyway, not a necessity.

[J G Miller]

On Sun, 18 Nov 2012 15:17:32 +0000, The Natural Philosopher wrote:

> Fat finger causes null post.

Was that aphorism originally uttered by a Chinese philosopher?

[Jerry Peters]

Pascal Hambourg <boite-...@plouf.fr.eu.org> wrote:
> Hello,
>
> unruh a �crit :
>> On 2012-11-16, Peter Hanke <pete...@andres.net> wrote:
>>> Assume I define in file /etc/fstab a new line for an additional tmpfs filesystem with mount point /tmp
>>> (and mount it either manually of through reboot).
>>>
>>> Then I copy some file stuff onto this filesystem dir /tmp
>>>
>>> Where are these files physically placed? Does /tmp only exist in memory RAM or inside an existing partition (which ?) on hard disc?
>>
>> In memory Ram.
>
> No, in virtual memory, i.e. either in system RAM or swap space.

>
>> It disappears if the computer is switched off.
>
> Swap space contents does not disappear when the computer is switched
> off. Even RAM contents does not disappear immediately.

But the information about what each swap page represents disappears
when the ram loses its contents.

Jerry

[Tim Watts]

Not to mention the physical hardness required. Vehicle electrinics on mining
trucks is bad enough - but the hardening required for systems on rockets is
something else, especially WRT to vibration.

A ZX81 wouldn't survive the door being shut before the trundle out to the
launchpad ;->

--
Tim Watts Personal Blog: http://www.dionic.net/tim/

"History will be kind to me for I intend to write it."

[John Hasler]

Michael writes:
> I can't even remember if they had a computer on board Apollo 11. If
> it was, it would be very minor, likely following a sequence of
> instructions rather than actually computer.

Both the command module and the lander had computers. And yes, they
were "real" 16 bit computers, built with RTL integrated circuits.

> The shuttle's computer came later, but not that much later since the
> design was frozen long before it made it into space.

The Shuttle had five 32-bit computers of the 360 architecture. They
used TTL ICs and core memory.

> I guess it was the Grid, which certainly was better than the on board
> computer,

"Better"? Would you trust your life to one?

> I seem to recal something like core memory, and it wasn't that great a
> computer.

It was an _excellent_ computer. It was not a very "powerful" one.
--
John Hasler
jha...@newsguy.com
Dancing Horse Hill
Elmwood, WI USA

[unruh]

On 2012-11-18, Aragorn <str...@telenet.be.invalid> wrote:
>
> Oh, that's not surprising. Any recent cellphone - it doesn't even have
> to be a smartphone - has more computing power than my dad's pre-owned
> hard-disk-less [*] XT clone in 1990. ;-)

As I understand it, the Raspberry Pi is a cell phone cpu. And it is more
powerful than computers available 10 years ago (never mind 20)

[The Natural Philosopher]

I am not chinese
As far as I know..

[Robert Riches]

On 2012-11-18, John Hasler <jha...@newsguy.com> wrote:
> Michael writes:
>> I can't even remember if they had a computer on board Apollo 11. If
>> it was, it would be very minor, likely following a sequence of
>> instructions rather than actually computer.
>
> Both the command module and the lander had computers. And yes, they
> were "real" 16 bit computers, built with RTL integrated circuits.
>
>> The shuttle's computer came later, but not that much later since the
>> design was frozen long before it made it into space.
>
> The Shuttle had five 32-bit computers of the 360 architecture. They
> used TTL ICs and core memory.

And, core memory _does_ retain its contents without power--at
least that's what was in the literature I quoted in an 8th-grade
research paper for Miss Beuhler's English class.

I read of one practice on the shuttle that took advantage of the
non-volatile nature of core memory. Reportedly, after getting
into orbit, the crew would load the software and data for landing
the shuttle into the computers and power them down. That way, if
the tape readers failed, the bits were already in core.

--
Robert Riches
spamt...@jacob21819.net
(Yes, that is one of my email addresses.)

[John Hasler]

Robert Riches writes:
> I read of one practice on the shuttle that took advantage of the
> non-volatile nature of core memory. Reportedly, after getting into
> orbit, the crew would load the software and data for landing the
> shuttle into the computers and power them down. That way, if the tape
> readers failed, the bits were already in core.

With core memory computers it was routine to simply shut them off with
every expectation that they would just pick up where they left off when
powered up again.