On 06/22/08 12:25 pm, Cork Soaker wrote:
> Given that a lot of hardware is now coming with SSDs, and we all know
> Windows will wear one of those little buggers out faster than anything
> else, [...]
>
Really? "Wear out" a solid state device? Perhaps you can expound further
how Windows manages to do such a thing?

--
jmm (hyphen) list (at) sohnen-moe (dot) com
(Remove .AXSPAMGN for email)

At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:

>
> On 06/22/08 12:25 pm, Cork Soaker wrote:
> > Given that a lot of hardware is now coming with SSDs, and we all know
> > Windows will wear one of those little buggers out faster than anything
> > else, [...]
> >
> Really? "Wear out" a solid state device? Perhaps you can expound further
> how Windows manages to do such a thing?

If the "Solid State Drive" is implemented with an EEPROM, then it is
possible -- EEPROMs have a limited number of re-write cycles.

>

--
Robert Heller -- Get the Deepwoods Software FireFox Toolbar!
Deepwoods Software -- Linux Installation and Administration
http://www.deepsoft.com/ -- Web Hosting, with CGI and Database
heller@deepsoft.com -- Contract Programming: C/C++, Tcl/Tk

Robert Heller wrote:
> At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:
>
>>
>> On 06/22/08 12:25 pm, Cork Soaker wrote:
>> > Given that a lot of hardware is now coming with SSDs, and we all know
>> > Windows will wear one of those little buggers out faster than anything
>> > else, [...]
>> >
>> Really? "Wear out" a solid state device? Perhaps you can expound further
>> how Windows manages to do such a thing?
>
> If the "Solid State Drive" is implemented with an EEPROM, then it is
> possible -- EEPROMs have a limited number of re-write cycles.

As do "FLASH" rom. Sometimes as low as 100,000 writes.
Better not make a swap partition or swap file on those things...

--
| spike1@freenet.co.uk | "I'm alive!!! I can touch! I can taste! |
| Andrew Halliwell BSc | I can SMELL!!! KRYTEN!!! Unpack Rachel and |
| in | get out the puncture repair kit!" |
| Computer Science | Arnold Judas Rimmer- Red Dwarf |

Andrew Halliwell writes:

> Robert Heller wrote:
>> At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:
>>
>>>
>>> On 06/22/08 12:25 pm, Cork Soaker wrote:
>>> > Given that a lot of hardware is now coming with SSDs, and we all know
>>> > Windows will wear one of those little buggers out faster than anything
>>> > else, [...]
>>> >
>>> Really? "Wear out" a solid state device? Perhaps you can expound further
>>> how Windows manages to do such a thing?
>>
>> If the "Solid State Drive" is implemented with an EEPROM, then it is
>> possible -- EEPROMs have a limited number of re-write cycles.
>
> As do "FLASH" rom. Sometimes as low as 100,000 writes.
> Better not make a swap partition or swap file on those things...

A flash device without a wear-levelling controller will wear out in a
very short time if used with a normal filesystem. Filesystems like
jffs have wear-levelling built in, so they are suitable for such
devices. I would assume that flash-based storage devices intended as
hard drive replacements have suitable controllers.

--
Måns Rullgård
mans@mansr.com

I demand that Cork Soaker may or may not have written...

> Given that a lot of hardware is now coming with SSDs, and we all know
> Windows will wear one of those little buggers out faster than anything
> else, is it (already) possible to set options that will only write back
> to the disk when it needs to, preferably at shutdown only?

# echo 1 >/proc/sys/vm/laptop_mode

http://www.kernel.org/doc/Documentation/laptops/

> I know there are some options to reduce the amount of writes to a disk,
> but how long can a write be avoided?

At best, for as long as there's memory available in which to buffer the data.

--
| Darren Salt | linux or ds at | nr. Ashington, | Toon
| RISC OS, Linux | youmustbejoking,demon,co,uk | Northumberland | Army
| Let's keep the pound sterling

Man and wife make one fool.

In article ,
Cork Soaker wrote:
> Given that a lot of hardware is now coming with SSDs, and we all know
> Windows will wear one of those little buggers out faster than anything
> else, is it (already) possible to set options that will only write back
> to the disk when it needs to, preferably at shutdown only?

If that's on, and you crash, you're hosed.

--
-eben QebWenE01R@vTerYizUonI.nOetP royalty.mine.nu:81
> A: It's annoying as hell
> Q: Why do most people hate top-posting? -- Lots42 The Library Avenger
http://www.fscked.co.uk/writing/top-posting-cuss.html

Måns Rullgård wrote:
> A flash device without a wear-levelling controller will wear out in a
> very short time if used with a normal filesystem. Filesystems like
> jffs have wear-levelling built in, so they are suitable for such
> devices. I would assume that flash-based storage devices intended as
> hard drive replacements have suitable controllers.

Even if the controllers do spread the load across as much of the FLASH as
possible, put a swap on there and it'll be worn out in less than a year,
most likely.
--
| spike1@freenet.co.uk | Windows95 (noun): 32 bit extensions and a |
| | graphical shell for a 16 bit patch to an 8 bit |
| Andrew Halliwell BSc | operating system originally coded for a 4 bit |
| in |microprocessor, written by a 2 bit company, that|
| Computer Science | can't stand 1 bit of competition. |

On 2008-06-23, Måns Rullgård wrote:
> Andrew Halliwell writes:
>
>> Robert Heller wrote:
>>> At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:
>>>
>>>>
>>>> On 06/22/08 12:25 pm, Cork Soaker wrote:
>>>> > Given that a lot of hardware is now coming with SSDs, and we all know
>>>> > Windows will wear one of those little buggers out faster than anything
>>>> > else, [...]
>>>> >
>>>> Really? "Wear out" a solid state device? Perhaps you can expound further
>>>> how Windows manages to do such a thing?
>>>
>>> If the "Solid State Drive" is implemented with an EEPROM, then it is
>>> possible -- EEPROMs have a limited number of re-write cycles.
>>
>> As do "FLASH" rom. Sometimes as low as 100,000 writes.
>> Better not make a swap partition or swap file on those things...
>
> A flash device without a wear-levelling controller will wear out in a
> very short time if used with a normal filesystem. Filesystems like
> jffs have wear-levelling built in, so they are suitable for such
> devices. I would assume that flash-based storage devices intended as
> hard drive replacements have suitable controllers.
>

I believe that all "SSDs" have wear leveling built in. I am waiting
for their price to come down and am very anxious ot use them,
especially for databases.
--
Due to extreme spam originating from Google Groups, and their inattention
to spammers, I and many others block all articles originating
from Google Groups. If you want your postings to be seen by
more readers you will need to find a different means of
posting on Usenet.
http://improve-usenet.org/

Jim Moe wrote:
> On 06/22/08 12:25 pm, Cork Soaker wrote:
>> Given that a lot of hardware is now coming with SSDs, and we all know
>> Windows will wear one of those little buggers out faster than anything
>> else, [...]
>>
> Really? "Wear out" a solid state device? Perhaps you can expound further
> how Windows manages to do such a thing?
>


You don't know??

Hactar wrote:
> In article ,
> Cork Soaker wrote:
>> Given that a lot of hardware is now coming with SSDs, and we all know
>> Windows will wear one of those little buggers out faster than anything
>> else, is it (already) possible to set options that will only write back
>> to the disk when it needs to, preferably at shutdown only?
>
> If that's on, and you crash, you're hosed.
>

Yes true, but Linux (Ubuntu for me) seems stable enough. :-)

If you're saving documents, I would expect it to save the file directly
to the disk, so a crash should only really lose settings. That's the
theory anyway....

Andrew Halliwell wrote:
> Robert Heller wrote:
>> At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:
>>
>>> On 06/22/08 12:25 pm, Cork Soaker wrote:
>>>> Given that a lot of hardware is now coming with SSDs, and we all know
>>>> Windows will wear one of those little buggers out faster than anything
>>>> else, [...]
>>>>
>>> Really? "Wear out" a solid state device? Perhaps you can expound further
>>> how Windows manages to do such a thing?
>> If the "Solid State Drive" is implemented with an EEPROM, then it is
>> possible -- EEPROMs have a limited number of re-write cycles.
>
> As do "FLASH" rom. Sometimes as low as 100,000 writes.
> Better not make a swap partition or swap file on those things...

Note that "solid state drive" can, and sometimes does, also mean battery
backed ram.

Andrew Halliwell wrote:
> Måns Rullgård wrote:
>> A flash device without a wear-levelling controller will wear out in a
>> very short time if used with a normal filesystem. Filesystems like
>> jffs have wear-levelling built in, so they are suitable for such
>> devices. I would assume that flash-based storage devices intended as
>> hard drive replacements have suitable controllers.
>
> Even if the controllers do spread the load across as much of the FLASH as
> possible, put a swap on there and it'll be worn out in less than a year,
> most likely.

I've been using 2/4GB Flash Disk-on-Module drives for the OS
installation in servers for a few years now and not had any problems.
They're not the quickest things in the world, but most bits of the OS
that are needed regularly get cached in RAM anyway. I also very rarely
see much use of swap these days, even on machines that are heavily loaded.

Just acquired a SATA SSD to try out. Should be a bit quicker than the DoM's.

Matthew

In comp.os.linux.hardware Ignoramus31759 wrote:
| On 2008-06-23, M?ns Rullg?rd wrote:
|> Andrew Halliwell writes:
|>
|>> Robert Heller wrote:
|>>> At Sun, 22 Jun 2008 13:46:45 -0700 Jim Moe wrote:
|>>>
|>>>>
|>>>> On 06/22/08 12:25 pm, Cork Soaker wrote:
|>>>> > Given that a lot of hardware is now coming with SSDs, and we all know
|>>>> > Windows will wear one of those little buggers out faster than anything
|>>>> > else, [...]
|>>>> >
|>>>> Really? "Wear out" a solid state device? Perhaps you can expound further
|>>>> how Windows manages to do such a thing?
|>>>
|>>> If the "Solid State Drive" is implemented with an EEPROM, then it is
|>>> possible -- EEPROMs have a limited number of re-write cycles.
|>>
|>> As do "FLASH" rom. Sometimes as low as 100,000 writes.
|>> Better not make a swap partition or swap file on those things...
|>
|> A flash device without a wear-levelling controller will wear out in a
|> very short time if used with a normal filesystem. Filesystems like
|> jffs have wear-levelling built in, so they are suitable for such
|> devices. I would assume that flash-based storage devices intended as
|> hard drive replacements have suitable controllers.
|>
|
| I believe that all "SSDs" have wear leveling built in. I am waiting
| for their price to come down and am very anxious ot use them,
| especially for databases.

Just because there is wear leveling does not mean they are immune to wearing
out. What wear leveling does is spread the wear out among other segments
But the segments on a flash device can be fairly large as the flash devices
get larger. The unit of wear leveling will be the erase segment size. What
is important to note is that only a free queue of available segments gets to
participate. Segments that already have data will not be erased. And the
number of additional segments that are not part of the addressable device are
not all that many. So writing over one small space many times spreads the
wear over just a small portion of the whole flash device. Writing over larger
areas increases the spread by the area increase.

If you are really going to write a lot, an extra layer of wear leveling might
help. The effect it will have is decreasing the available space and increasing
the available free queue. That would be doing JFFS2 over a device that already
has hardware wear leveling.

--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
| Usenet from these places, find another Usenet provider ASAP. |
| Phil Howard KA9WGN (email for humans: first name in lower case at ipal.net) |