8:13 a.m.
It appears that among some kernel maintainers there's an opinion that
the hibernate (suspend to disk) capability is of insufficient interest
to users to justify the difficulty of maintenance. See
https://bugzilla.redhat.com/show_bug.cgi?id=781749
Will anyone for whom hibernate is important please add a
note to the bug report, so that the maintainers will know that their
efforts are much appreciated?
Thanks.
7:06 p.m.
On 06 Mar 2012 at 18:07:24, Marko Vojinovic wrote:
On Tuesday, 6. March 2012. 17.42.41 Anthony R Fletcher wrote:
> If you do a fresh boot you lose your current state....and that is
> valuable. So hibernate (for long periods of down time, for example for a
> long flight) is very useful.
What happened to the "save session" stuff that should be provided by all
relevant desktop environments? AFAIK, that can (and should) be used to save
the current state of your desktop across a reboot. Is there some aspect of
session-saving that doesn't give you back your desktop "state" in the way
you
left it on logout?
I mean, it should open the same apps, keep them on same desktops, etc. I
thought the concept of a "session" was invented precisely for this purpose.
Using hibernate to achieve the same effect is possible, but should not be
necessary, right?
That might work with some GUI programs but if the user has a bunch of
xterms open with some work in each then it just isn't going to get saved
correctly. Another example is some long running calculation that you
don't want to lose but does have a checkpoint feature. This would be
true with many other work flows as well. Why fight with each individual
application when you can just hibernate the machine.
Anthony.
4:42 p.m.
On 06 Mar 2012 at 17:35:22, Reindl Harald wrote:
Am 06.03.2012 22:49, schrieb Rex Dieter:
> Geoffrey Leach wrote:
>
>> It appears that among some kernel maintainers there's an opinion that
>> the hibernate (suspend to disk) capability is of insufficient interest
>> to users to justify the difficulty of maintenance.
>
> It's not an issue about users' interest at all. Obviously its a useful
> feature.
on machines these days?
you can guess how long it takes dump 16 GB to disk and load
it compared with a full boot between 10 and 30 seconds (30
seconds with a LOT of services like mail, www, mysql...)
If you do a fresh boot you lose your current state....and that is
valuable. So hibernate (for long periods of down time, for example for a
long flight) is very useful.
4:42 p.m.
On 03/07/2012 04:05 AM, Reindl Harald wrote:
Am 06.03.2012 22:49, schrieb Rex Dieter:
> Geoffrey Leach wrote:
>
>> It appears that among some kernel maintainers there's an opinion that
>> the hibernate (suspend to disk) capability is of insufficient interest
>> to users to justify the difficulty of maintenance.
>
> It's not an issue about users' interest at all. Obviously its a useful
> feature.
on machines these days?
Sure. Depends on what you use them for.
Rahul
4:35 p.m.
Am 06.03.2012 22:49, schrieb Rex Dieter:
Geoffrey Leach wrote:
> It appears that among some kernel maintainers there's an opinion that
> the hibernate (suspend to disk) capability is of insufficient interest
> to users to justify the difficulty of maintenance.
It's not an issue about users' interest at all. Obviously its a useful
feature.
on machines these days?
you can guess how long it takes dump 16 GB to disk and load
it compared with a full boot between 10 and 30 seconds (30
seconds with a LOT of services like mail, www, mysql...)
7:03 p.m.
On 03/06/2012 03:07 PM, Marko Vojinovic wrote:
On Tuesday, 6. March 2012. 17.42.41 Anthony R Fletcher wrote:
> If you do a fresh boot you lose your current state....and that is
> valuable. So hibernate (for long periods of down time, for example for a
> long flight) is very useful.
What happened to the "save session" stuff that should be provided by all
relevant desktop environments? AFAIK, that can (and should) be used to save
the current state of your desktop across a reboot. Is there some aspect of
session-saving that doesn't give you back your desktop "state" in the way
you
left it on logout?
I mean, it should open the same apps, keep them on same desktops, etc. I
thought the concept of a "session" was invented precisely for this purpose.
Using hibernate to achieve the same effect is possible, but should not be
necessary, right?
No. There is no way for the desktop to, say, know just where the hell
you were if you were editing a document or viewing a movie or anything
of that type of transient nature. The _application_ knows where it was,
the desktop only knows where the windows were and how your desktop
looked.
I use suspend during the trips between my office and home (short),
hibernate to disk if it'll be an extended trip (like a plane trip).
Hibernate is an extremely useful tool. Do not confuse it with "save
session". They're very different beasties and meant for different
things.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks(a)alldigital.com -
- AIM/Skype: therps2 ICQ: 22643734 Yahoo: origrps2 -
- -
- Squawk! Pieces of Seven! Pieces of Seven! Parity Error! -
----------------------------------------------------------------------
5:07 p.m.
On Tuesday, 6. March 2012. 17.42.41 Anthony R Fletcher wrote:
If you do a fresh boot you lose your current state....and that is
valuable. So hibernate (for long periods of down time, for example for a
long flight) is very useful.
What happened to the "save session" stuff that should be provided by all
relevant desktop environments? AFAIK, that can (and should) be used to save
the current state of your desktop across a reboot. Is there some aspect of
session-saving that doesn't give you back your desktop "state" in the way
you
left it on logout?
I mean, it should open the same apps, keep them on same desktops, etc. I
thought the concept of a "session" was invented precisely for this purpose.
Using hibernate to achieve the same effect is possible, but should not be
necessary, right?
Best, :-)
Marko
5:16 p.m.
On Wed, Mar 7, 2012 at 12:07 AM, Marko Vojinovic <vvmarko(a)gmail.com> wrote:
On Tuesday, 6. March 2012. 17.42.41 Anthony R Fletcher wrote:
> If you do a fresh boot you lose your current state....and that is
> valuable. So hibernate (for long periods of down time, for example for a
> long flight) is very useful.
What happened to the "save session" stuff that should be provided by all
relevant desktop environments? AFAIK, that can (and should) be used to save
the current state of your desktop across a reboot. Is there some aspect of
session-saving that doesn't give you back your desktop "state" in the way
you
left it on logout?
+1
I mean, it should open the same apps, keep them on same desktops, etc. I
thought the concept of a "session" was invented precisely for this purpose.
Using hibernate to achieve the same effect is possible, but should not be
necessary, right?
Best, :-)
Marko
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
--
Best,
Christopher Svanefalk
1:32 p.m.
On Thu, 2012-05-17 at 23:52 +0800, Steve Underwood wrote:
Quite right. Only someone who never uses a laptop could think
hibernate
and suspend are no longer needed.
I use hibernate daily on my desktop too. For various reasons, recovering
a complete session, for me, can take quite a while. The time it takes to
boot the system and log in is only a minority of this time. Hibernation
saves me time and drudge work every single day.
--Greg
4:57 a.m.
10 years ago i lived more than 3 years without a job
and in this time you have not much money
And if you were doing that in most of Western Europe/Canada/etc you would
have been receivign vastly more than someone in many other countries
working 50hours a week.
I would question your maths on disk costs and electricity pricing too. To
an extent its an open question given there's a human work cost involved
but even a reasonably efficient modern machine running 24 x 7 is using a
fair amount of kWh with associated costs.
Suspend cuts a lot of this and in some cases may well be a better option.
Virtual machines also change the patterns as you can migrate your working
environment between systems.
Alan
3:46 p.m.
Timothy Murphy wrote:
You are assuming that because your colleague's HP cost €800
it is of "higher quality" than my daughter's under €300 Asus netbook,
where by "higher quality" you apparently mean "will last longer".
From my - very long but not broad - experience
there is little or no correlation between the price of a computer
or computer component and how long it lasts.
Absolutely. In some cases a high-end piece of equipment (say, a graphics
card) is going to run hotter and be more power-hungry and be more likely
to fail than a low-end one, not least because there will be more fans to
fail.
Fortunately, this is no longer the case, but ten to fifteen years ago,
unless you bought carefully, when you bought a cheap desktop you got a
cheap processor on a cheap motherboard with a dodgy chipset, and if you
spent a bit more you got a faster version of the same processor on the
same dodgy motherboard.¹
I bet that Asus netbook is based on an Atom processor and an Intel
chipset, and whatever else you might say about Intel (and who doesn’t?),
they still (mostly) remember the twenty-year-old Pentium lesson that if
it’s got their name on it, it ought to be reliable.
A couple of related quotes from people who can say it better than I can:
All Hardware Sucks, and yesteryear’s professional-grade stuff?
Consumer-grade $#!=€ where they’re worried about losing their thin
profit margin to a few too many returns can be a lot more reliable
than professional gear where they were drooling about gouging you on
support. Caveat emptor.
– Anthony de Boer
Tandy went bust about a decade ago due to high prices, bad customer
service, and punting crap that nobody wanted. Maplin saw that market
niche and repositioned themselves into it, without asking the
obvious question as to why said niche didn’t have competition.
– Peter Corlett
¹ And back then, if the CPU fan did fail, it was more likely to be fatal
for faster, hotter processors.
--
E-mail: james@ | "This was, apparently, beyond her ken.
aprilcottage.co.uk | So far beyond her ken that she was well into barbie
| territory."
| -- J. D. Baldwin
8:24 a.m.
Reindl Harald wrote:
again: swap may be usefull if you have too few RAM
on modern machines starting with 8-16 GB this is esotheric
I agree. With the advent of x86_64 and dirt cheap RAM swap
files/partitions should be phased out. Hibernate and suspend are no
longer necessary or helpful functions either (on machines sold today).
1:24 p.m.
Out of curiosity, I thought a few years ago that the future would be
hibernate implemented by kexec... And then no one else mentioned it
anymore.
Does anyone know what happened? Lack of interest or techical issues?
Maybe this would be a good time to revive the interest on it, if possible?
--
Pedro
11:27 a.m.
Reindl Harald wrote:
but for only write some mails and web-brwosing i can not see why 10
seconds
more or less are making the difference and as cheaper/slower your
hardware is as longer wake up from suspend takes
I'm not sure exactly what you are saying.
I don't care if my laptop takes 10 seconds more or less to start.
I was simply pointing out that contrary to what you said,
waking from hibernation takes less time than re-booting,
at least on all the laptops I've had where hibernation was an option.
The main advantage of hibernation for me
is that the machine returns in exactly the same state I left it,
with the same web-sites open in the browser,
and with virtual desktops open at the same places.
If I forget to close a file I am editing, which I often do,
it is still open when I re-start.
--
Timothy Murphy
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin
7:06 a.m.
Am 18.05.2012 13:25, schrieb Timothy Murphy:
You mention having machines with 16GB RAM.
How exactly is this used?
I just checked my 2 servers;
the one here (I'm in Italy) has 5GB RAM, 4GB of which is free
("Mem: 4829188k total, 847540k used, 3981648k free").
The server in Ireland has 4GB RAM, with 500MB free
("Mem: 3793928k total, 3209440k used, 584488k free")
on servers:
* MySQL with large InnoDB-Buffers (http://www.dbmail.org/)
* Web-Applications where you expect more than a few users
* Bytecode-Caches and Query-Cache for MySQL
* VMware Workstation in Headless-Mode on Homeservers for isolation / testing
* VMware ESXi on production hosts
* many other services i mentioned before
__________________________________
on workstations:
* Eclipse / ZendStudio (it is very happy to have >= 700 MB for it's own)
* Desktop
* Firefox
* Thunderbird
* VMware Workstation
* OpenOffice
* OS-Cache
* all the services using in production on serveral veritual machines for testing
half a hour after i started my machine it is using around 3GB active
yes, my focus is on heavy use
but for only write some mails and web-brwosing i can not see why 10 seconds
more or less are making the difference and as cheaper/slower your hardware
is as longer wake up from suspend takes
6:25 a.m.
Reindl Harald wrote:
but that was never a reason to buy low quality
You are assuming that because your colleague's HP cost €800
it is of "higher quality" than my daughter's under €300 Asus netbook,
where by "higher quality" you apparently mean "will last longer".
From my - very long but not broad - experience
there is little
or no correlation between the price of a computer
or computer component and how long it lasts.
I'm genuinely puzzled by the vastly different prices of computers
which seem to me to offer more or less the same facilities.
Most people, I think, have fairly modest expectations for their laptops;
they want to browse, send e-mail, read newsgroups or forums,
perhaps look at movies.
As for home servers, the number of services required is very small.
I have two HP MicroServers (in Ireland and Italy)
running dhcpd, httpd, openVPN, openLDAP, dovecot/IMAP, mySQL
and not much more (shorewall, fail2ban, etc).
Each cost under €200, I think, because of HPs crazy pricing system.
But they don't seem to find the load too much for them.
I see servers that cost 5 times as much advertised as "home servers",
and I'm genuinely puzzled what additional or better facilities they offer.
You mention having machines with 16GB RAM.
How exactly is this used?
I just checked my 2 servers;
the one here (I'm in Italy) has 5GB RAM, 4GB of which is free
("Mem: 4829188k total, 847540k used, 3981648k free").
The server in Ireland has 4GB RAM, with 500MB free
("Mem: 3793928k total, 3209440k used, 584488k free").
--
Timothy Murphy
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin
5:11 a.m.
Am 18.05.2012 11:57, schrieb Alan Cox:
I would question your maths on disk costs and electricity pricing
too. To
an extent its an open question given there's a human work cost involved
but even a reasonably efficient modern machine running 24 x 7 is using a
fair amount of kWh with associated costs.
60-70W is not much, this are only a few € each year
keep in mind that this is not more than a bulb you
forgot to switch off in the morining before you go
to work
keep also in mind that this bublb is not that much , your
energy costs, they are caused by your refrigerator, your
dishwasher, microwave oven, your cooker and such things
and not by one machine with 60-70W while this is not only
the machine
the UPS is measuring switches, modem, printer and the SIP-phone
too which are not off if you suspend only the PC
so the power of this one machine alone does not matter and if it
is running it can do scheduled jobs like backups from machines
into a virtual container and typical cleanup-jobs you do not want
to run if you are working
4:44 a.m.
Am 18.05.2012 10:18, schrieb Heinz Diehl:
On 18.05.2012, Reindl Harald wrote:
> buying a cheaper machine than 800 € results usually
> in buy much more machines in a relative short term
Maybe for you, but not for folks who hardly can afford a low end
machine. Not having a computer is no longer an alternative these days
either.
10 years ago i lived more than 3 years without a job
and in this time you have not much money
but that was never a reason to buy low quality
You are quite obviously not clear over that most of the people
using computers do not know at all what a power cycle is..
this is not a technical argument
educate them
4:41 a.m.
Am 18.05.2012 10:36, schrieb Timothy Murphy:
You've repeated this several times,
so I thought I'd test it on my laptop,
a Thinkpad T60 running Fedora-16/KDE.
I did each test twice.
Hibernate (ie suspend to disk) and shutdown
both took the same time, 18-20 seconds.
Waking from hibernation took 40-41 seconds.
Cold boot + login took 72-73 seconds
cold boot: 25 seconds
login: 10 seconds
not counting the time to enter login and wallet paswords
5 seconds for me
bother different passwords with 19 chars
so i am around 40 seconds too
additionally i have to log out and switch
to a console if i have used my home-machine
between to sync back changes consistently
many programs do not like sync back their profiles
when they are running
The laptop was unusable for a further 30 seconds,
due I presume to disk activity, mainly virtuoso-t and
firefox.
well, i have the same unuseable expierience waking up
machines from suspend on a changed location because
dns-caches are holding LAN addresses from the company
and freezed connections since the IP's are no longer
available
3:36 a.m.
Reindl Harald wrote:
waking up from suspend to disk takes much longer as a cold
start
You've repeated this several times,
so I thought I'd test it on my laptop,
a Thinkpad T60 running Fedora-16/KDE.
I did each test twice.
Hibernate (ie suspend to disk) and shutdown
both took the same time, 18-20 seconds.
Waking from hibernation took 40-41 seconds.
Cold boot + login took 72-73 seconds,
not counting the time to enter login and wallet paswords.
The laptop was unusable for a further 30 seconds,
due I presume to disk activity, mainly virtuoso-t and
firefox.
--
Timothy Murphy
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin
3:18 a.m.
On 18.05.2012, Reindl Harald wrote:
buying a cheaper machine than 800 € results usually
in buy much more machines in a relative short term
Maybe for you, but not for folks who hardly can afford a low end
machine. Not having a computer is no longer an alternative these days
either.
cheaper than disks killed by power-cycles and my
current homeserver eats around 50-60 watt at
normal operations - you can run this machine
some years before costs for electricity beats
costs for a single disk dying due power cycles
You are quite obviously not clear over that most of the people using
computers do not know at all what a power cycle is..
5:01 p.m.
On 17May2012 22:33, Reindl Harald <h.reindl(a)thelounge.net> wrote:
| i said only suspned-to-disk is unuseable if you have
| a modern machine with >= 16 GB RAM or a notebook with
| 8 RAM and slow notebook-disks
If all this happens _after_ I close the laptop lid I often don't care.
Admittedly, I'm using a machine that sleeps/suspends - it ony has to
come back from a real hibernate if the laptop battery goes flat before I
reopen it.
--
Cameron Simpson <cs(a)zip.com.au> DoD#743
http://www.cskk.ezoshosting.com/cs/
Life is pleasant. Death is peaceful. It's the transition that's troublesome.
- Isaac Asimov
4:57 p.m.
On 05/17/2012 02:13 PM, Reindl Harald wrote:
Am 17.05.2012 23:08, schrieb Joe Zeff:
> If you assume he's talking about his own specific needs, not the
> general case, what he writes makes much more sense.
uhm it is clear and logical for me that i speak about
my needs and expierience in the last 10 years
for who else could i?
It may be clear and logical to you, but the way you write makes it look
like yours is The One True Way and that you expect everybody else to do
things your way. This, BTW, is why you end up in so many arguments;
your writing style is both inflexible and argumentative. You might want
to work on that.
4:56 p.m.
On 17May2012 11:33, Chris Adams <cmadams(a)hiwaay.net> wrote:
| Windows 7 also has a nice thing that Linux does not: a hybrid mode
| between suspend and hibernate. With that, when you suspend, it goes
| through the hiberate steps (writing what it needs to disk), but then
| puts the system into suspend. If the battery runs dead or the power
| fails, you can restore from hibernate. If the system hasn't lost power,
| when you restore it is just a wake from suspend.
The MacBooks also.
I should point out that this all happens after I shut the lid. So to a
degree it does not matter how long it takes to save to disc, because I
am not sitting around waiting for it to happen.
So:
shut lid
Mac runs hibernate disc write (sits with LED on steady still)
...then suspends (LED starts its slow "I'm asleep" pulse)
Cheers,
--
Cameron Simpson <cs(a)zip.com.au> DoD#743
http://www.cskk.ezoshosting.com/cs/
Inspector Clay's dead. Murdered. And somebody is responsible!
- _Plan 9 From Outer Space_
4:51 p.m.
On 17May2012 23:52, Steve Underwood <steveu(a)coppice.org> wrote:
| On 05/17/2012 11:15 PM, Tim wrote:
| > On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
| >> Hibernate and suspend are no longer necessary or helpful functions
| >> either (on machines sold today).
| > I suspend on my Laptop, all the time. Quite apart from the speed issue,
| > it's handy to be able to halt and resume, everything.
| Quite right. Only someone who never uses a laptop could think hibernate
| and suspend are no longer needed.
Yep. I'm often on a Macbook (yes, not Linux - that's on the home server)
and one of the real joys is, when I run out of time somewhere (the
train arrives, etc) just shut the lid and stuff it into into my bag. No
planning or special arrangements. It suspends to RAM, and if the battery
really does go flat later, restores from the hard drive when I attach
real power.
On a laptop this is a great boon.
--
Cameron Simpson <cs(a)zip.com.au> DoD#743
http://www.cskk.ezoshosting.com/cs/
NOTWORK: n. A network when it is acting flaky. Origin (?) IBM.
- Hackers' Dictionary
4:48 p.m.
On 17May2012 08:30, Dave Ihnat <dihnat(a)dminet.com> wrote:
| It would make sense to offer the option to disable swap, hibernate,
| standby, etc.; but don't remove it from the system. To do so is to say
| *you* know better than *I* do how capable my hardware may be; that's
| positively Microsoft- or Apple-ish.
To be fair, Apple at least do know "how capable my hardware may be"
since they make their own.
But yes, end users should have the say, even if the OS makes default
decisions.
--
Cameron Simpson <cs(a)zip.com.au> DoD#743
http://www.cskk.ezoshosting.com/cs/
If it can't be turned off, it's not a feature. - Karl Heuer
4:31 p.m.
Am 17.05.2012 23:22, schrieb Timothy Murphy:
>> I don't consider Fedora suitable for a server.
>
> your opinion
and I suspect that of most people who have to make this choice.
> if you need a recent software stack and have to
> compile all things at your won while libraries
> are outdated you are going through hell
I find CentOS-6 has everything I need on a server.
I don't feel any urge to re-compile libraries.
I have complete confidence that RedHat will do this for me if necessary.
If I want to do anything fancy I do it on a laptop.
currently this may be true
version 6 is quite recent
but what in 5 years?
if you need newer PHP, MySQL, Postfix, DBmail, Apache
you have a problem even with compile it at your own
because you can not solve build-deps
caused by the large jumps you can usually not upgrade
a CentOS installation and having a lot of machines
and renstall them all few years is not funny
on the other hand i have here 20 fedora servers which
are beoming more over the time based on the same
goldenmaster with a identical coresystem installed with
F9 and currently on F16, they are all hardly tuned
and optimizd for exatcly their needs and replace hardware
is a single click to move them uninterrupted
dist-upgrades are a topic in virtual environments
as i installed them all you needed VMware Tools
for basic paravirtualized hardware which is in
the meantime in the upstream.kernel
another example now for phyical hardware:
kernel 3.4 will support replace a RADI10 with bigger
disks and expand the RAID volumes - you will never
see this in CentOS6, with fedora in a few weeks it is
possible
4:22 p.m.
Reindl Harald wrote:
> Michael Cronenworth wrote:
Actually I said what you are quoting.
> You are talking nonsense if by "modern" you mean
"recent".
>
> My daughter just bought an Asus laptop (1015BX) with 1GB RAM installed,
> and a maximum 2GB RAM installable.
here is the question why buying crap these days?
my co-worker bought last year a notebook with 4 GB RAM
normal HP machine around 800 € a year ago
My daughter's Asus notebook cost under €300, I think around €275.
It appears to serve her needs perfectly.
> I don't consider Fedora suitable for a server.
your opinion
and I suspect that of most people who have to make this choice.
if you need a recent software stack and have to
compile all things at your won while libraries
are outdated you are going through hell
I find CentOS-6 has everything I need on a server.
I don't feel any urge to re-compile libraries.
I have complete confidence that RedHat will do this for me if necessary.
If I want to do anything fancy I do it on a laptop.
--
Timothy Murphy
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin
4:18 p.m.
On Thu, May 17, 2012 at 3:24 PM, Michael Cronenworth <mike(a)cchtml.com> wrote:
Reindl Harald wrote:
> again: swap may be usefull if you have too few RAM
> on modern machines starting with 8-16 GB this is esotheric
I agree. With the advent of x86_64 and dirt cheap RAM swap
files/partitions should be phased out. Hibernate and suspend are no
longer necessary or helpful functions either (on machines sold today).
--
Suspend no longer helpful? It is extremely helpful. I run a new
desktop and I constantly use it. It incomparably much better than
shutting down every time you have to leave for a longer period...not
to mention that you can "boot" in no time right back into your
previous desktop session this way. I never turn off the machine or
reboot unless I have to.
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
--
Best,
Christopher Svanefalk
4:16 p.m.
On Thu, May 17, 2012 at 02:08:43PM -0700, Joe Zeff wrote:
On 05/17/2012 01:54 PM, Darryl L. Pierce wrote:
>Then that's not a usable system is it? I'm not talking about the base
>boot speed but the boot speed with everything running. I have none of the
>ones you list below running on my system exception httpd and my boot
>time is 10s of seconds, as I said before.
At this point, I think he's trying to explain why hibernate isn't
practical for him, not why it's not a good idea in general.
I don't know, maybe that's the case. But he doesn't seem to talk about
his own use case when he seems to be telling others that their scenarios
are "meaningless".
Alas,
as I've pointed out to him recently, his "tone of typer" makes it
look like he's saying, "I don't use it so it's no good for
anybody."
If you assume he's talking about his own specific needs, not the
general case, what he writes makes much more sense. Let's not get
into Yet Another Flame War over this.
I like to think I'm good about walking away when it's obvious the
discussion's going nowhere. :)
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
4:15 p.m.
On Thu, May 17, 2012 at 11:01:34PM +0200, Reindl Harald wrote:
>> you did read the "a lot of services"?
>> disable them and you are around 8-10 seconds on F16
>
> Then that's not a usable system is it? I'm not talking about the base
> boot speed but the boot speed with everything running. I have none of the
> ones you list below running on my system exception httpd and my boot
> time is 10s of seconds, as I said before.
well, that exactly is what i mean by "suspend to disk is meaningless"
the coffe machine takes longer than 10 seconds
I'm sorry, but I haven't a clue what you're getting at here. I'm telling
you that my system takes longer to cold boot than it does to resume from
suspend, which goes against your claim that cold boots are always
faster. They are in fact not since a reboot requires a lot more to occur
than just loading an active image from disk.
keep in mind we are speeking about normal disks
* in a few years most typical machines will have SSD
* the running apps of the sessions are started by the DE
None of this matters since, in my case at least, I'm not talking about
the login process when I say cold boot. Restoring from suspend is faster
than a boot, excluding the login.
from this moment on suspend is meaningless from the
point of performance compared to a cold boot and more
or less a thing of taste
I think this is what others are pointing out regarding your tone. You
keep saying "is meaningless" but fail to realize that it's not. That
_you_ find no meaning does not mean _I_ don't find meaning (as in
usefulness) from such a feature.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
4:08 p.m.
On 05/17/2012 01:54 PM, Darryl L. Pierce wrote:
Then that's not a usable system is it? I'm not talking about
the base
boot speed but the boot speed with everything running. I have none of the
ones you list below running on my system exception httpd and my boot
time is 10s of seconds, as I said before.
At this point, I think he's trying to explain why hibernate isn't
practical for him, not why it's not a good idea in general. Alas, as
I've pointed out to him recently, his "tone of typer" makes it look like
he's saying, "I don't use it so it's no good for anybody." If you
assume he's talking about his own specific needs, not the general case,
what he writes makes much more sense. Let's not get into Yet Another
Flame War over this.
4:01 p.m.
Am 17.05.2012 22:54, schrieb Darryl L. Pierce:
On Thu, May 17, 2012 at 10:44:42PM +0200, Reindl Harald wrote:
> you did read the "a lot of services"?
> disable them and you are around 8-10 seconds on F16
Then that's not a usable system is it? I'm not talking about the base
boot speed but the boot speed with everything running. I have none of the
ones you list below running on my system exception httpd and my boot
time is 10s of seconds, as I said before.
well, that exactly is what i mean by "suspend to disk is meaningless"
the coffe machine takes longer than 10 seconds
keep in mind we are speeking about normal disks
* in a few years most typical machines will have SSD
* the running apps of the sessions are started by the DE
from this moment on suspend is meaningless from the
point of performance compared to a cold boot and more
or less a thing of taste
3:54 p.m.
On Thu, May 17, 2012 at 10:44:42PM +0200, Reindl Harald wrote:
Am 17.05.2012 22:37, schrieb Darryl L. Pierce:
>> currently 25 seconds including a lot of services
>> not used on a typical end-user machine
>
> Not so quickly for me. Granted my swap and home partitions are
> encrypted, but the password entry is hard a second or two during the
> boot process.
you did read the "a lot of services"?
disable them and you are around 8-10 seconds on F16
Then that's not a usable system is it? I'm not talking about the base
boot speed but the boot speed with everything running. I have none of the
ones you list below running on my system exception httpd and my boot
time is 10s of seconds, as I said before.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
3:44 p.m.
Am 17.05.2012 22:37, schrieb Darryl L. Pierce:
> currently 25 seconds including a lot of services
> not used on a typical end-user machine
Not so quickly for me. Granted my swap and home partitions are
encrypted, but the password entry is hard a second or two during the
boot process.
you did read the "a lot of services"?
disable them and you are around 8-10 seconds on F16
apcupsd.service
dbmail-imapd.service
dbmail-lmtpd.service
dbmail-timsieved.service
dovecot.service
httpd.service
mpd.service
mpdscribble.service
mysqld.service
named.service
pure-ftpd.service
replication.service
smb.service
vmware.service
> depends on your environment
>
> if you are connected to a lot of LAn services and wake
> up the machine on another location where they are all
> not available or have different IPs it is not funny
Still, that's an issue separate from hibernating or suspending a
machine that should be handled similar to any other network interruption
scenario
it is not a simple "network interruption" if you are leaving a LAN
with all servcices like web/mail/dns/fileserver and wake up the
machine with a lot of open connections to destinations which
are no longer existing with their old IP
3:40 p.m.
On Thu, May 17, 2012 at 10:33:00PM +0200, Reindl Harald wrote:
i said only suspned-to-disk is unuseable if you have
a modern machine with >= 16 GB RAM or a notebook with
8 RAM and slow notebook-disks
someone may find it nice to have his previous desktop
state, but it is NOT faster than a cold boot
I might point out that many people keep a lot of programs active--for
instance, a development IDE, several browser windows, debugger, chat
server, maybe word processor, etc. For them, it's certainly a lot faster
to restore from a standby or hibernation state with all these programs
already loaded, active, and positioned than to reboot, then open
all their applications.
The upshot? Everyone uses their machine differently. What's slow for you
isn't for them.
Cheers,
--
Dave Ihnat
dihnat(a)dminet.com
3:37 p.m.
On Thu, May 17, 2012 at 10:27:55PM +0200, Reindl Harald wrote:
>> i live in the world where someone starts his work in the
>> morining and powers on his computer once each day and
>> have all other machines running 365/7/24
>>
>> waking up from suspend to disk takes much longer as a cold start
>
> Can you provide some data to back this up? When I suspend my laptop it
> is far, far quicker to restore than a cold boot.
yes
Okay, I was asking you to actually provide that data, Reindl. ;)
> A suspend-to-usable operation is on the order of seconds
reading 16 GB RAm image in seconds?
not with slow disks
Yep. My laptop (Lenovo Thinkpad W510+) goes from suspend to running in a
few seconds. Hibernate's been broken in F17 for me so no stats there ATM
(which is why I'm interested in this thread).
> A cold boot is 10s of seconds.
currently 25 seconds including a lot of services
not used on a typical end-user machine
Not so quickly for me. Granted my swap and home partitions are
encrypted, but the password entry is hard a second or two during the
boot process.
>> and even if this is not interesting my expierience with
applications
>> and services having open network connections is that it sucks if they
>> are woken up in another network
>
> The machine should be able to handle it like any other interruption to
> networking (network down, switching APs, etc.). If it doesn't then
> that's a separate problem to be solved.
depends on your environment
if you are connected to a lot of LAn services and wake
up the machine on another location where they are all
not available or have different IPs it is not funny
Still, that's an issue separate from hibernating or suspending a
machine that should be handled similar to any other network interruption
scenario.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
3:33 p.m.
Am 17.05.2012 22:29, schrieb Joe Zeff:
On 05/17/2012 01:09 PM, Reindl Harald wrote:
> i live in the world where someone starts his work in the
> morining and powers on his computer once each day and
> have all other machines running 365/7/24
Therefor, hibernate is wrong *for you.* Not everybody lives in your world; some people
find that *for them*
hibernate is simpler or faster than shutdown and reboot. (As it happens, I'm not one
of them.) This is yet
another case where your way isn't the only right way. Deal with it.
WHERE did i say that my way is the only right?
i said only suspned-to-disk is unuseable if you have
a modern machine with >= 16 GB RAM or a notebook with
8 RAM and slow notebook-disks
someone may find it nice to have his previous desktop
state, but it is NOT faster than a cold boot
3:29 p.m.
On 05/17/2012 01:09 PM, Reindl Harald wrote:
i live in the world where someone starts his work in the
morining and powers on his computer once each day and
have all other machines running 365/7/24
Therefor, hibernate is wrong *for you.* Not everybody lives in your
world; some people find that *for them* hibernate is simpler or faster
than shutdown and reboot. (As it happens, I'm not one of them.) This
is yet another case where your way isn't the only right way. Deal with it.
3:27 p.m.
Am 17.05.2012 22:23, schrieb Darryl L. Pierce:
On Thu, May 17, 2012 at 10:09:36PM +0200, Reindl Harald wrote:
> i live in the world where someone starts his work in the
> morining and powers on his computer once each day and
> have all other machines running 365/7/24
>
> waking up from suspend to disk takes much longer as a cold start
Can you provide some data to back this up? When I suspend my laptop it
is far, far quicker to restore than a cold boot.
yes
A suspend-to-usable operation is on the order of seconds
reading 16 GB RAm image in seconds?
not with slow disks
A cold boot is 10s of seconds.
currently 25 seconds including a lot of services
not used on a typical end-user machine
> and even if this is not interesting my expierience with
applications
> and services having open network connections is that it sucks if they
> are woken up in another network
The machine should be able to handle it like any other interruption to
networking (network down, switching APs, etc.). If it doesn't then
that's a separate problem to be solved.
depends on your environment
if you are connected to a lot of LAn services and wake
up the machine on another location where they are all
not available or have different IPs it is not funny
3:23 p.m.
On Thu, May 17, 2012 at 10:09:36PM +0200, Reindl Harald wrote:
i live in the world where someone starts his work in the
morining and powers on his computer once each day and
have all other machines running 365/7/24
waking up from suspend to disk takes much longer as a cold start
Can you provide some data to back this up? When I suspend my laptop it
is far, far quicker to restore than a cold boot. A suspend-to-usable
operation is on the order of seconds. A cold boot is 10s of seconds.
and even if this is not interesting my expierience with applications
and services having open network connections is that it sucks if they
are woken up in another network
The machine should be able to handle it like any other interruption to
networking (network down, switching APs, etc.). If it doesn't then
that's a separate problem to be solved.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
3:22 p.m.
Am 17.05.2012 22:19, schrieb Alan Cox:
On Thu, 17 May 2012 22:09:36 +0200
Reindl Harald <h.reindl(a)thelounge.net> wrote:
>> My daughter just bought an Asus laptop (1015BX) with 1GB RAM installed,
>> and a maximum 2GB RAM installable.
>
> here is the question why buying crap these days?
Its a pretty typical low end netbook configuration. See "finanical
crisis". See "relative buying power".
So its a very very common configuration in much of the World.
buying a cheaper machine than 800 € results usually
in buy much more machines in a relative short term
> i live in the world where someone starts his work in the
> morining and powers on his computer once each day and
> have all other machines running 365/7/24
Where presumably electricity charges are nearly free
cheaper than disks killed by power-cycles and my
current homeserver eats around 50-60 watt at
normal operations - you can run this machine
some years before costs for electricity beats
costs for a single disk dying due power cycles
3:19 p.m.
On Thu, 17 May 2012 22:09:36 +0200
Reindl Harald <h.reindl(a)thelounge.net> wrote:
> Michael Cronenworth wrote:
> You are talking nonsense if by "modern" you mean "recent".
>
> My daughter just bought an Asus laptop (1015BX) with 1GB RAM installed,
> and a maximum 2GB RAM installable.
here is the question why buying crap these days?
Its a pretty typical low end netbook configuration. See "finanical
crisis". See "relative buying power".
So its a very very common configuration in much of the World.
i live in the world where someone starts his work in the
morining and powers on his computer once each day and
have all other machines running 365/7/24
Where presumably electricity charges are nearly free.
Alan
3:09 p.m.
Michael Cronenworth wrote:
You are talking nonsense if by "modern" you mean "recent".
My daughter just bought an Asus laptop (1015BX) with 1GB RAM installed,
and a maximum 2GB RAM installable.
here is the question why buying crap these days?
my co-worker bought last year a notebook with 4 GB RAM
normal HP machine around 800 € a year ago
I don't consider Fedora suitable for a server.
your opinion
if you need a recent software stack and have to
compile all things at your won while libraries
are outdated you are going through hell
I suspect you live in a different world to the rest of us,
and should be wary of giving advice on what should or should not
be included in Fedora.
Ps I guess this should be directed at Reindl Harald, not you
i live in the world where someone starts his work in the
morining and powers on his computer once each day and
have all other machines running 365/7/24
waking up from suspend to disk takes much longer as a cold start
and even if this is not interesting my expierience with applications
and services having open network connections is that it sucks if they
are woken up in another network
12:52 p.m.
On Thu, May 17, 2012 at 06:08:34PM +0200, Reindl Harald wrote:
Am 17.05.2012 17:52, schrieb Steve Underwood:
> On 05/17/2012 11:15 PM, Tim wrote:
>> On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
>>> Hibernate and suspend are no longer necessary or helpful functions
>>> either (on machines sold today).
>> I suspend on my Laptop, all the time. Quite apart from the speed issue,
>> it's handy to be able to halt and resume, everything.
>
> Quite right. Only someone who never uses a laptop could think hibernate and suspend
are no longer needed.
sure?
i used a laptop from 2003 until 2011 as my main working machine
all the day and never came to the idea write a 6 GB to a slow
mobile-disk and load it the next time instead simply shutdown/boot
And I used to do it all the time. There's no reason to shutdown and
reboot a system unless there's a kernel update, so being able to suspend
or hibernate a system and then restore at a later date is a convenience.
I only don't do it now because hibernate doesn't work for me with F17.
--
Darryl L. Pierce, Sr. Software Engineer @ Red Hat, Inc.
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
7:46 a.m.
On 17.5.2012 15:36, Reindl Harald wrote:
Am 17.05.2012 14:29, schrieb Joel Rees:
> Do you understand the reason you still set up swap, even though your
> entire workload working set fits into RAM?
there is no single reason if you have neough RAM
RAM is better used for disk cache than for idling processes. If a
process does nothing it can be swapped out and the memory used where needed.
--
Don't plan any hasty moves. You'll be evicted soon anyway.
11:39 a.m.
Once upon a time, Steven Stern <subscribed-lists(a)sterndata.com> said:
Wouldn't it be A Wonderful Thing if a Fedora laptop (or even a
Windows
laptop) could suspend and restart as effortlessly as a MacBook? When I
close the lid on my MacBook, I am completely confident it will work when
I pop it back open.
My Thinkpad with Fedora "just works" for suspend. Hibernate works
correctly as well; I only use it sometimes, but I've occasionally been
suprised by a hibernate resume instead of suspend resume (when the
battery gets low in suspend, the system wakes up and hibernates).
--
Chris Adams <cmadams(a)hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.
11:42 a.m.
Am 17.05.2012 18:33, schrieb Chris Adams:
Once upon a time, Reindl Harald <h.reindl(a)thelounge.net> said:
> sure?
Yes.
> i used a laptop from 2003 until 2011 as my main working machine
> all the day and never came to the idea write a 6 GB to a slow
> mobile-disk and load it the next time instead simply shutdown/boot
Okay, so you don't want to use hibernate. Do you also not even suspend?
as "simply shutdown/boot" -> no!
beside from the fact that i do not too often power on / off my
machines and the start in th emorining happens by drink coffee
for backup-reasons
i have all the time one 24/7 machine at home and EVERY TIME i
shut down my working machine it does a rsync to the home machine
of all my data - thunderbird/firefox as one example are not
really happy if you copy their sqlite files while they are
open -> menas your backup is inconsistent
yes, i am aware the this is not the workload of many people
but many people have lost some piece of data in the past
i never did because it is impossible that the is no recent backup
11:33 a.m.
On 05/17/2012 11:08 AM, Reindl Harald wrote:
Am 17.05.2012 17:52, schrieb Steve Underwood:
> On 05/17/2012 11:15 PM, Tim wrote:
>> On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
>>> Hibernate and suspend are no longer necessary or helpful functions
>>> either (on machines sold today).
>> I suspend on my Laptop, all the time. Quite apart from the speed issue,
>> it's handy to be able to halt and resume, everything.
>
> Quite right. Only someone who never uses a laptop could think hibernate and suspend
are no longer needed.
sure?
i used a laptop from 2003 until 2011 as my main working machine
all the day and never came to the idea write a 6 GB to a slow
mobile-disk and load it the next time instead simply shutdown/boot
Wouldn't it be A Wonderful Thing if a Fedora laptop (or even a Windows
laptop) could suspend and restart as effortlessly as a MacBook? When I
close the lid on my MacBook, I am completely confident it will work when
I pop it back open.
--
-- Steve
11:33 a.m.
Once upon a time, Reindl Harald <h.reindl(a)thelounge.net> said:
sure?
Yes.
i used a laptop from 2003 until 2011 as my main working machine
all the day and never came to the idea write a 6 GB to a slow
mobile-disk and load it the next time instead simply shutdown/boot
Okay, so you don't want to use hibernate. Do you also not even suspend?
I've got a fast SSD in my notebook, and it hibernates faster than I
would boot up, restart all my programs, and reopen any files I had when
I last used it.
Windows 7 also has a nice thing that Linux does not: a hybrid mode
between suspend and hibernate. With that, when you suspend, it goes
through the hiberate steps (writing what it needs to disk), but then
puts the system into suspend. If the battery runs dead or the power
fails, you can restore from hibernate. If the system hasn't lost power,
when you restore it is just a wake from suspend.
--
Chris Adams <cmadams(a)hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.
11:08 a.m.
Am 17.05.2012 17:52, schrieb Steve Underwood:
On 05/17/2012 11:15 PM, Tim wrote:
> On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
>> Hibernate and suspend are no longer necessary or helpful functions
>> either (on machines sold today).
> I suspend on my Laptop, all the time. Quite apart from the speed issue,
> it's handy to be able to halt and resume, everything.
Quite right. Only someone who never uses a laptop could think hibernate and suspend are
no longer needed.
sure?
i used a laptop from 2003 until 2011 as my main working machine
all the day and never came to the idea write a 6 GB to a slow
mobile-disk and load it the next time instead simply shutdown/boot
10:57 a.m.
Once upon a time, Steve Underwood <steveu(a)coppice.org> said:
On 05/17/2012 11:15 PM, Tim wrote:
>On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
>>Hibernate and suspend are no longer necessary or helpful functions
>>either (on machines sold today).
>I suspend on my Laptop, all the time. Quite apart from the speed issue,
>it's handy to be able to halt and resume, everything.
Quite right. Only someone who never uses a laptop could think hibernate
and suspend are no longer needed.
I suspend desktops that I occasionally need to access remotely, my HTPC
(when the TV is off), and more. Suspend is hardly only for laptops.
The desktops that I suspend are supposed to wake up and hibernate on
power failure (before the UPS dies), but I haven't tested that in a
while.
--
Chris Adams <cmadams(a)hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.
10:52 a.m.
On 05/17/2012 11:15 PM, Tim wrote:
On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
> Hibernate and suspend are no longer necessary or helpful functions
> either (on machines sold today).
I suspend on my Laptop, all the time. Quite apart from the speed issue,
it's handy to be able to halt and resume, everything.
Quite right. Only someone
who never uses a laptop could think hibernate
and suspend are no longer needed.
Steve
10:15 a.m.
On Thu, 2012-05-17 at 08:24 -0500, Michael Cronenworth wrote:
Hibernate and suspend are no longer necessary or helpful functions
either (on machines sold today).
I suspend on my Laptop, all the time. Quite apart from the speed issue,
it's handy to be able to halt and resume, everything.
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
9:04 a.m.
Michael Cronenworth wrote:
> again: swap may be usefull if you have too few RAM
> on modern machines starting with 8-16 GB this is esotheric
I agree. With the advent of x86_64 and dirt cheap RAM swap
files/partitions should be phased out. Hibernate and suspend are no
longer necessary or helpful functions either (on machines sold today).
You are talking nonsense if by "modern" you mean "recent".
My daughter just bought an Asus laptop (1015BX) with 1GB RAM installed,
and a maximum 2GB RAM installable.
My home server is an HP MicroServer with 4GB RAM.
I could install 8GB but see no reason to.
Incidentally, I'm running CentOS on it;
I don't consider Fedora suitable for a server.
No-one is forcing you to use hibernate.
As for it being offered as an option,
surely there are dozens of options that you never use?
I suspect you live in a different world to the rest of us,
and should be wary of giving advice on what should or should not
be included in Fedora.
Ps I guess this should be directed at Reindl Harald, not you.
--
Timothy Murphy
e-mail: gayleard /at/ eircom.net
tel: +353-86-2336090, +353-1-2842366
s-mail: School of Mathematics, Trinity College Dublin
7:36 a.m.
Am 17.05.2012 14:29, schrieb Joel Rees:
> you can guess how long it takes dump 16 GB to disk and load
> it
Guess, ....
Or calculate?
calculate it
it takes way too long
it may acceptable on machines with real fast RAID10
but they are booting also much faster and are up in 10-15 seconds
5 seconds for login and a few seconds for session-restore
Do you understand the reason you still set up swap, even though your
entire workload working set fits into RAM?
there is no single reason if you have neough RAM
> compared with a full boot between 10 and 30 seconds (30
> seconds with a LOT of services like mail, www, mysql...)
The Gimp?
GIMP starts in around 5 seconds on recent machines
8:38 a.m.
Am 17.05.2012 15:30, schrieb Dave Ihnat:
On Thu, May 17, 2012 at 08:24:50AM -0500, Michael Cronenworth wrote:
> I agree. With the advent of x86_64 and dirt cheap RAM swap
> files/partitions should be phased out. Hibernate and suspend are no
> longer necessary or helpful functions either (on machines sold today).
With all due respect, do recall that many who are installing Linux are
doing so on older or recycled machines.
yes, but i spoke from RECENT hardware
Also, Linux is used in a wide
variety of environments, not just desktops.
i know
i am using Fedora as router and for > 20 servers
It would make sense to offer the option to disable swap, hibernate,
standby, etc.; but don't remove it from the system. To do so is to say
*you* know better than *I* do how capable my hardware may be; that's
positively Microsoft- or Apple-ish.
i did not say anything opposite
my original answer was to the proven wrong "Do you understand the reason
you still set up swap, even though your entire workload working set
fits into RAM?" and "RAM is better used for disk cache than for idling
processes. If a process does nothing it can be swapped out and the memory
used where needed"
proven by > 20 servers and some large desktop systems with a lot of
mixed usecases over many years using no single byte of swap most
of the time and having only some hundret MB swap-FILE and sysctl
with "vm.swappiness = 1"
the expierience after all this years is if your system starts to
swap anyhting performance is going down dramatically, ever on any
machine
8:30 a.m.
On Thu, May 17, 2012 at 08:24:50AM -0500, Michael Cronenworth wrote:
I agree. With the advent of x86_64 and dirt cheap RAM swap
files/partitions should be phased out. Hibernate and suspend are no
longer necessary or helpful functions either (on machines sold today).
With all due respect, do recall that many who are installing Linux are
doing so on older or recycled machines. Also, Linux is used in a wide
variety of environments, not just desktops.
It would make sense to offer the option to disable swap, hibernate,
standby, etc.; but don't remove it from the system. To do so is to say
*you* know better than *I* do how capable my hardware may be; that's
positively Microsoft- or Apple-ish.
Cheers,
--
Dave Ihnat
dihnat(a)dminet.com
8:03 a.m.
Am 17.05.2012 14:46, schrieb Jari Fredriksson:
On 17.5.2012 15:36, Reindl Harald wrote:
>
>
> Am 17.05.2012 14:29, schrieb Joel Rees:
>> Do you understand the reason you still set up swap, even though your
>> entire workload working set fits into RAM?
>
> there is no single reason if you have neough RAM
>
RAM is better used for disk cache than for idling processes. If a
process does nothing it can be swapped out and the memory used where needed
says who?
if you have running applications than for a reason
swap them out for cache POSSIBLY re-needed disk blocks
will cause delays and re-read from disk
on a typical workload with 8-16 GB RAm there is no sible reason
for swap, especially on workstations but mostly also on
servers if they have enough RAM
and i am speaking here about a server hosting some hundret domains
as example which has 7000 mysql tables, no single byte swapped
and a load-avg around 0.8
again: swap may be usefull if you have too few RAM
on modern machines starting with 8-16 GB this is esotheric
7:29 a.m.
Where do we get these recruits?
On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
Am 06.03.2012 22:49, schrieb Rex Dieter:
> Geoffrey Leach wrote:
>
>> It appears that among some kernel maintainers there's an opinion that
>> the hibernate (suspend to disk) capability is of insufficient interest
>> to users to justify the difficulty of maintenance.
>
> It's not an issue about users' interest at all. Obviously its a useful
> feature.
on machines these days?
Netbook?
you can guess how long it takes dump 16 GB to disk and load
it
Guess, ....
Or calculate?
Do you understand the reason you still set up swap, even though your
entire workload working set fits into RAM?
compared with a full boot between 10 and 30 seconds (30
seconds with a LOT of services like mail, www, mysql...)
The Gimp?
--
Joel Rees
10:29 a.m.
Stop this! (please)
2012/5/28 Joel Rees <joel.rees(a)gmail.com>:
On Sat, May 26, 2012 at 5:20 PM, Reindl Harald
<h.reindl(a)thelounge.net> wrote:
>
>
> Am 26.05.2012 07:45, schrieb Joel Rees:
>>>> Do you understand the reason you still set up swap, even though your
>>>> entire workload working set fits into RAM?
>>>
>>> there is no single reason if you have enough RAM
>>
>> In an ideal world, RAM would not consume energy.
>>
>> This is a real world, what energy I have on the train is a small
>> Lithium ion battery.
>
> why in the world do you wake up this thread after weeks again?
The after three weeks part is that I have a day job and a night job,
and trying to keep up with the mailing list is not top priority for
me.
The wake it up part, I guess, is that I really, really think hibernate
is important.
And technical accuracy is also important to me.
You contribute some good things to the list, Riendl, but your field of
vision seems severely limited to your own working set. And some of the
things you say (nothing but greedy RAM policies in any Unix-like OS?
No reason to use a low-power netbook running Fedora on the train?) are
just plain wrong.
--
Joel Rees
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
8:23 a.m.
On Sat, May 26, 2012 at 5:20 PM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
Am 26.05.2012 07:45, schrieb Joel Rees:
>>> Do you understand the reason you still set up swap, even though your
>>> entire workload working set fits into RAM?
>>
>> there is no single reason if you have enough RAM
>
> In an ideal world, RAM would not consume energy.
>
> This is a real world, what energy I have on the train is a small
> Lithium ion battery.
why in the world do you wake up this thread after weeks again?
The after three weeks part is that I have a day job and a night job,
and trying to keep up with the mailing list is not top priority for
me.
The wake it up part, I guess, is that I really, really think hibernate
is important.
And technical accuracy is also important to me.
You contribute some good things to the list, Riendl, but your field of
vision seems severely limited to your own working set. And some of the
things you say (nothing but greedy RAM policies in any Unix-like OS?
No reason to use a low-power netbook running Fedora on the train?) are
just plain wrong.
--
Joel Rees
4:18 p.m.
please guys, start your private flamefest and leave us alone
2012/5/26 Reindl Harald <h.reindl(a)thelounge.net>:
Am 26.05.2012 13:50, schrieb Heinz Diehl:
> On 26.05.2012, Reindl Harald wrote:
>
>>> Why do you think it's wrong?
>
>> it is wrong
>
> This discussion ends here for me. Your claims and statements totally
> lack any evidence, which means that it's a waste of time to read
> any further.
please do so
the discussion ended for me on 2012-03-07 and is re-started
multiple times by Joel Rees which has nothing better to do
than quote me and restart from begin
nobody can change the fact that i see no benefit in hibernate
and swap after 10 years in IT business so please Joel
leave us in peace with you wakeup of this discussion or at
least quote somebody other than me!
-------- Original-Nachricht --------
Betreff: Re: The death of Hibernate?
Datum: Thu, 17 May 2012 21:29:17 +0900
Von: Joel Rees <joel.rees(a)gmail.com>
Where do we get these recruits?
On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
-------- Original-Nachricht --------
Betreff: Re: The death of Hibernate?
Datum: Sat, 26 May 2012 14:45:29 +0900
Von: Joel Rees <joel.rees(a)gmail.com>
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
7:36 a.m.
Am 26.05.2012 13:50, schrieb Heinz Diehl:
On 26.05.2012, Reindl Harald wrote:
>> Why do you think it's wrong?
> it is wrong
This discussion ends here for me. Your claims and statements totally
lack any evidence, which means that it's a waste of time to read
any further.
please do so
the discussion ended for me on 2012-03-07 and is re-started
multiple times by Joel Rees which has nothing better to do
than quote me and restart from begin
nobody can change the fact that i see no benefit in hibernate
and swap after 10 years in IT business so please Joel
leave us in peace with you wakeup of this discussion or at
least quote somebody other than me!
-------- Original-Nachricht --------
Betreff: Re: The death of Hibernate?
Datum: Thu, 17 May 2012 21:29:17 +0900
Von: Joel Rees <joel.rees(a)gmail.com>
Where do we get these recruits?
On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
-------- Original-Nachricht --------
Betreff: Re: The death of Hibernate?
Datum: Sat, 26 May 2012 14:45:29 +0900
Von: Joel Rees <joel.rees(a)gmail.com>
6:21 a.m.
Am 26.05.2012 11:41, schrieb Heinz Diehl:
Millions of small notebooks with 1 GB RAM have been sold,
because they are small, able to run many hours solely on batteri,
and because a lot of people don't have the money to buy something more
expensive. In other words: they are quite common, and most of them run
Linux just fine.
have yu considered that a böeeding edge distribution
maybe wrong for the oldest and cheapest hardware?
there are distributions specialized for such hardware
> the argumentation of the crazy guy above was that RAm is
> better used for caching as for unused running applications
> which is completly wrong
Why do you think it's wrong?
it is wrong
There's more than just "one world"
there is only one world: swapping sucks performance
depends entirely on what you want to do with your machine. Maybe, the
"crazy
guy above" is not as crazy as you claim he is:
not really
"My point is that decreasing the tendency of the kernel to swap
stuff
out is wrong. You really don't want hundreds of megabytes of BloatyApp's
untouched memory floating about in the machine. Get it out on the
disk, use the memory for something useful." (Andrew Morton)
my point is that increasing the tendency of the kernel to swap stuff
out is wrong, you really don't want hundrets of megabytes of
bloaty diskblocks untouched in memory and swap out applications
use the memory for applications instead of wait around a minute if
yo change the focus back to a application the OS swapped out
4:41 a.m.
On 26.05.2012, Reindl Harald wrote:
and even if - machines with 1 GB RAM are loughable
these days since i remember that a yum-upgrade was
killed with a OOM om a virtual machine witout GUI
Millions of small notebooks with 1 GB RAM have been sold,
because they are small, able to run many hours solely on batteri,
and because a lot of people don't have the money to buy something more
expensive. In other words: they are quite common, and most of them run
Linux just fine.
the argumentation of the crazy guy above was that RAm is
better used for caching as for unused running applications
which is completly wrong
Why do you think it's wrong? There's more than just "one world", it
depends entirely on what you want to do with your machine. Maybe, the "crazy
guy above" is not as crazy as you claim he is:
"My point is that decreasing the tendency of the kernel to swap stuff
out is wrong. You really don't want hundreds of megabytes of BloatyApp's
untouched memory floating about in the machine. Get it out on the
disk, use the memory for something useful." (Andrew Morton)
3:20 a.m.
Am 26.05.2012 07:45, schrieb Joel Rees:
>> Do you understand the reason you still set up swap, even
though your
>> entire workload working set fits into RAM?
>
> there is no single reason if you have enough RAM
In an ideal world, RAM would not consume energy.
This is a real world, what energy I have on the train is a small
Lithium ion battery.
why in the world do you wake up this thread after weeks again?
hibernate does not inerest me but i can not see people argumentating t
echnically wrong
it does not change anything if yopu are swap out all your
applications because you will ALWYAS have 100% RAM used
on any unix-like system
and even if - machines with 1 GB RAM are loughable
these days since i remember that a yum-upgrade was
killed with a OOM om a virtual machine witout GUI
braindead developers cosuming ressources for nothing
this is not new remember in 2003 where i had aa machine
with 192 MB RAM running Win2000, Photoshop, CorelDrwaw
and VMware with a linux-guest and these days a f**ing
updater needs much more
__________________________________
the argumentation of the crazy guy above was that RAm is
better used for caching as for unused running applications
which is completly wrong
where in the world is the benefit having applications swapped
out for chaching the blocks of a 2 GB vieo file moved around
and not touching for weeks
1:55 p.m.
On Sat, May 19, 2012 at 8:41 PM, Joe Zeff <joe(a)zeff.us> wrote:
On 05/19/2012 05:48 AM, Christopher Svanefalk wrote:
>
> As for desktops, it should be noted of course that the one-piece iMac
> is a very different concept than a standard case-monitor desktop.
But not, by any means, a new one. To me it looks like a modern version of a
typical CP/M machine of the late '70s or early '80s.
True, only that Apple has done it so good that, for the average
consumer, it might just feel like the most novel and useful thing on
earth.
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
--
Best,
Christopher Svanefalk
1:41 p.m.
On 05/19/2012 05:48 AM, Christopher Svanefalk wrote:
As for desktops, it should be noted of course that the one-piece
iMac
is a very different concept than a standard case-monitor desktop.
But not, by any means, a new one. To me it looks like a modern version
of a typical CP/M machine of the late '70s or early '80s.
7:30 a.m.
Am 19.05.2012 13:47, schrieb Dave Ihnat:
Actually, there's a bit of misconception here. Yes, you can find
machines
with marginal configurations that may have the same processor as a Macbook
Pro that appear to be very much cheaper.
But they have less or slower memory, less capable/slower support chipsets,
slower/smaller hard drives, less capable video cards, "home" versions
of Windows (if you go that route), etc. By the time you configure an
Intel-based machine to meet the same performance specs as an equivalent
Apple machine, the difference shrinks to a couple of hundred dollars
or less. THAT is the legitimate "Apple Tax", not the hundreds or
thousands people claim.
this may be PARTLY true for a notebook, on the other hand i have naver
seen so many displays or graphics cards die as in macbooks or currently
screens from the iMacs
but not for a desktop or any professional computer
the machine below:
http://www8.hp.com/de/de/products/desktops/product-detail.html?oid=5096737
* 4x2 TB Disk as RAID10
* 16 GB RAM
* Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz
1600 € a year ago
you will not find any compareable machine from Apple
the iMacs are only toys and the XEON/ECC of a MacPro is not needed
___________________________
00:00.0 Host bridge: Intel Corporation 2nd Generation Core Processor Family DRAM
Controller (rev 09)
00:01.0 PCI bridge: Intel Corporation Xeon E3-1200/2nd Generation Core Processor Family
PCI Express Root Port (rev 09)
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family
Integrated Graphics
Controller (rev 09)
00:19.0 Ethernet controller: Intel Corporation 82579LM Gigabit Network Connection (rev
04)
00:1a.0 USB Controller: Intel Corporation 6 Series/C200 Series Chipset Family USB Enhanced
Host Controller #2 (rev 04)
00:1b.0 Audio device: Intel Corporation 6 Series/C200 Series Chipset Family High
Definition Audio Controller (rev 04)
00:1c.0 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express Root
Port 1 (rev b4)
00:1c.4 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express Root
Port 5 (rev b4)
00:1c.6 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express Root
Port 7 (rev b4)
00:1c.7 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express Root
Port 8 (rev b4)
00:1d.0 USB Controller: Intel Corporation 6 Series/C200 Series Chipset Family USB Enhanced
Host Controller #1 (rev 04)
00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev a4)
00:1f.0 ISA bridge: Intel Corporation Q67 Express Chipset Family LPC Controller (rev 04)
00:1f.2 SATA controller: Intel Corporation 6 Series/C200 Series Chipset Family 6 port SATA
AHCI Controller (rev 04)
00:1f.3 SMBus: Intel Corporation 6 Series/C200 Series Chipset Family SMBus Controller (rev
04)
01:00.0 Ethernet controller: Intel Corporation 82574L Gigabit Network Connection
02:00.0 USB Controller: NEC Corporation uPD720200 USB 3.0 Host Controller (rev 03)
03:00.0 Network controller: Atheros Communications Inc. AR5008 Wireless Network Adapter
(rev 01)
6:47 a.m.
On Sat, May 19, 2012 at 10:34:31AM +0200, Christopher Svanefalk wrote:
Seems to me Apple has been doing this masterfully for years now. How
much better is a $2000 Macbook Pro as opposed to a HP laptop with the
same processor?
Actually, there's a bit of misconception here. Yes, you can find machines
with marginal configurations that may have the same processor as a Macbook
Pro that appear to be very much cheaper.
But they have less or slower memory, less capable/slower support chipsets,
slower/smaller hard drives, less capable video cards, "home" versions
of Windows (if you go that route), etc. By the time you configure an
Intel-based machine to meet the same performance specs as an equivalent
Apple machine, the difference shrinks to a couple of hundred dollars
or less. THAT is the legitimate "Apple Tax", not the hundreds or
thousands people claim.
Cheers,
--
Dave Ihnat
dihnat(a)dminet.com
3:34 a.m.
On Sat, May 19, 2012 at 9:29 AM, Heinz Diehl <htd(a)fritha.org> wrote:
On 19.05.2012, James Wilkinson wrote:
> I bet that Asus netbook is based on an Atom processor and an Intel
> chipset, and whatever else you might say about Intel (and who doesn’t?),
> they still (mostly) remember the twenty-year-old Pentium lesson that if
> it’s got their name on it, it ought to be reliable.
Q: According to Intel, the Pentium conforms to the IEEE standards 754
and 854 for floating point arithmetic. If you fly in aircraft
designed using a Pentium, what is the correct pronunciation of
"IEEE"?
A: Aaaaaaaiiiiiiiiieeeeeeeeeeeee!
[SCNR!]
Seriously, I think you're totally right. Pricing these days seems to
follow marketing strategies rather than reflecting quality. Most of
the components are the same, and "fashion" has a big impact on how
much they want to have for a particular machine. E.g. if you want
something that is very slim and looks sharp, you'll have to pay double
the price of a machine containing the same core components, but is
standard sized.
Seems to me Apple has been doing this masterfully for years now. How
much better is a $2000 Macbook Pro as opposed to a HP laptop with the
same processor?
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
--
Best,
Christopher Svanefalk
2:29 a.m.
On 19.05.2012, James Wilkinson wrote:
I bet that Asus netbook is based on an Atom processor and an Intel
chipset, and whatever else you might say about Intel (and who doesn’t?),
they still (mostly) remember the twenty-year-old Pentium lesson that if
it’s got their name on it, it ought to be reliable.
Q: According to Intel, the Pentium conforms to the IEEE standards 754
and 854 for floating point arithmetic. If you fly in aircraft
designed using a Pentium, what is the correct pronunciation of
"IEEE"?
A: Aaaaaaaiiiiiiiiieeeeeeeeeeeee!
[SCNR!]
Seriously, I think you're totally right. Pricing these days seems to
follow marketing strategies rather than reflecting quality. Most of
the components are the same, and "fashion" has a big impact on how
much they want to have for a particular machine. E.g. if you want
something that is very slim and looks sharp, you'll have to pay double
the price of a machine containing the same core components, but is
standard sized.
5:33 p.m.
Thats horrible news - Microsoft will switch to hibernation as the
default way of booting with Windows8 while on linux its so broken that
it will be disabled by default. If only the Ubuntu guys would
contribute a little bit more, instead of coding their own
"replacements" for various things.
I always wated to use hibernation, but unfourtunatly my Toshiba Tecra
A8 had to really buggy BIOS.
3:49 p.m.
Geoffrey Leach wrote:
It appears that among some kernel maintainers there's an opinion
that
the hibernate (suspend to disk) capability is of insufficient interest
to users to justify the difficulty of maintenance.
It's not an issue about users' interest at all. Obviously its a useful
feature.
What it *is* (imo),
* there are a not-so-insignificant number of hibernation-related kernel bugs
* no one (esp upstream) with the interest and ability to fix said bugs
-- rex
1:39 a.m.
On Tue, Mar 6, 2012 at 10:13 PM, Geoffrey Leach <geoff(a)hughes.net> wrote:
It appears that among some kernel maintainers there's an opinion
that
the hibernate (suspend to disk) capability is of insufficient interest
to users to justify the difficulty of maintenance.
Personally, hibernation is quite a useful feature.
I hate to open the same documents every time when I start working.
12:45 a.m.
On Thu, May 17, 2012 at 9:36 PM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
Am 17.05.2012 14:29, schrieb Joel Rees:
>> you can guess how long it takes dump 16 GB to disk and load
>> it
>
> Guess, ....
>
> Or calculate?
calculate it
I'd like to see your calculations, although I got a clue from your
comment about slow disks.
it takes way too long
There is a reason for that, other than your harping about how booting
fits your needs better than suspending or hibernating.
it may acceptable on machines with real fast RAID10
but they are booting also much faster and are up in 10-15 seconds
5 seconds for login and a few seconds for session-restore
Well, on my lenovo S100 ideapaddy or whatever this piece of junk is
called, I have problems getting a good session restore on reboot. It
can wedge xfce4 so that I have to use some workaround to get a shell
and use the --replace option. Bugs. Cheap hardware that I could
actually buy, instead of dream about, to replace the iBook that died
permanently a bit before Christmas.
But, as far as which is faster, I have 1G of RAM. I don't run database
servers or web servers on it. Go ahead, guess or calculate.
I mostly use it on the train for e-mail, text editing, and some
translation work that I'm hoping will stretch the paycheck through the
last week of the month before payday. Sometimes I use it at work to do
things that are not allowed on the computers at work. (Yeah, "tacit"
permission that could bite me, so I only use it for emergencies.) An
Android phone or tablet would be more appropriate, but I could not
afford that, period. Economic realities, here.
Boot time is not what I want to do on the train. Suspend takes maybe
five seconds to sleep, about the same to wake up. Hibernate takes,
actually, not much more, ten seconds max. Cold boot takes somewhere
between thirty seconds and more than a minute.
> Do you understand the reason you still set up swap, even though
your
> entire workload working set fits into RAM?
there is no single reason if you have neough RAM
In an ideal world, RAM would not consume energy.
This is a real world, what energy I have on the train is a small
Lithium ion battery. Better than a set of NiCd cells, but still quite
limited. And even "just" 1G of RAM consumes quite a bit less suspended
than running, and even less hibernated. Enough that I can suspend,
shut the lid, get off the train, work all morning, and my work state
doesn't disappear in a power-down before lunch. Leaving it running, it
might force power-down by the time I walk twenty minutes from the
station to work.
By the way, in an ideal world (my version), the netbook I'm carrying
would not be a lenovo Intel. It wouldn't even be a cold-fire or
ARM-based unit. It'd be running a port of Linux or one of the BSDs on
a swarm of Forth processors.
And nothing but the currently active apps would keep state in RAM,
which would save huge time on hibernate and restore, because you
wouldn't be trying to dump and restore the entire system RAM. (Maybe
there's a hint in there as to why your experience with hibernate seems
to have you thinking you don't want to do that.)
>> compared with a full boot between 10 and 30 seconds (30
>> seconds with a LOT of services like mail, www, mysql...)
>
> The Gimp?
GIMP starts in around 5 seconds on recent machines
Modulo your definition of recent. But you missed the point.
LibreOffice, Inkscape, whatever. If I power down the netbook before I
get off the train, I have to save whatever I'm working on and quit the
apps. (Yeah, actually, I've broken the gimp out and worked a little on
graphics on the train once or twice.) That takes time. Then, after I
boot back up, I have to open whatever it was I was working on, both
starting the apps and loading the documents.
With suspend or hibernate, yeah, it's safest to save, but I don't have
to quit. And I don't have to start the apps back up after booting back
up.
By the way, even at home, running the 40W netbook all day long would
cost about half of what we spend on lighting for the month. It adds
up, and in Japan, it's not as cheap as some other places.
I am glad you find you don't need hibernate or suspend.
But this thread started with someone talking about kernel devs that
want to get rid of suspend and hibernate. If they did remove those
from the kernel, I would need something to replace them.
--
Joel Rees
7:48 a.m.
On Sat, May 19, 2012 at 2:30 PM, Reindl Harald <h.reindl(a)thelounge.net> wrote:
Am 19.05.2012 13:47, schrieb Dave Ihnat:
> Actually, there's a bit of misconception here. Yes, you can find machines
> with marginal configurations that may have the same processor as a Macbook
> Pro that appear to be very much cheaper.
>
> But they have less or slower memory, less capable/slower support chipsets,
> slower/smaller hard drives, less capable video cards, "home" versions
> of Windows (if you go that route), etc. By the time you configure an
> Intel-based machine to meet the same performance specs as an equivalent
> Apple machine, the difference shrinks to a couple of hundred dollars
> or less. THAT is the legitimate "Apple Tax", not the hundreds or
> thousands people claim.
this may be PARTLY true for a notebook, on the other hand i have naver
seen so many displays or graphics cards die as in macbooks or currently
screens from the iMacs
but not for a desktop or any professional computer
the machine below:
http://www8.hp.com/de/de/products/desktops/product-detail.html?oid=5096737
* 4x2 TB Disk as RAID10
* 16 GB RAM
* Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz
1600 € a year ago
you will not find any compareable machine from Apple
the iMacs are only toys and the XEON/ECC of a MacPro is not needed
___________________________
00:00.0 Host bridge: Intel Corporation 2nd Generation Core Processor Family DRAM
Controller (rev 09)
00:01.0 PCI bridge: Intel Corporation Xeon E3-1200/2nd Generation Core Processor Family
PCI Express Root Port (rev 09)
00:02.0 VGA compatible controller: Intel Corporation 2nd Generation Core Processor Family
Integrated Graphics
Controller (rev 09)
00:19.0 Ethernet controller: Intel Corporation 82579LM Gigabit Network Connection (rev
04)
00:1a.0 USB Controller: Intel Corporation 6 Series/C200 Series Chipset Family USB
Enhanced Host Controller #2 (rev 04)
00:1b.0 Audio device: Intel Corporation 6 Series/C200 Series Chipset Family High
Definition Audio Controller (rev 04)
00:1c.0 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express
Root Port 1 (rev b4)
00:1c.4 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express
Root Port 5 (rev b4)
00:1c.6 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express
Root Port 7 (rev b4)
00:1c.7 PCI bridge: Intel Corporation 6 Series/C200 Series Chipset Family PCI Express
Root Port 8 (rev b4)
00:1d.0 USB Controller: Intel Corporation 6 Series/C200 Series Chipset Family USB
Enhanced Host Controller #1 (rev 04)
00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev a4)
00:1f.0 ISA bridge: Intel Corporation Q67 Express Chipset Family LPC Controller (rev 04)
00:1f.2 SATA controller: Intel Corporation 6 Series/C200 Series Chipset Family 6 port
SATA AHCI Controller (rev 04)
00:1f.3 SMBus: Intel Corporation 6 Series/C200 Series Chipset Family SMBus Controller
(rev 04)
01:00.0 Ethernet controller: Intel Corporation 82574L Gigabit Network Connection
02:00.0 USB Controller: NEC Corporation uPD720200 USB 3.0 Host Controller (rev 03)
03:00.0 Network controller: Atheros Communications Inc. AR5008 Wireless Network Adapter
(rev 01)
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Have a question? Ask away: http://ask.fedoraproject.org
Good points from both posters above.
I guess what has to be factored in is the fact that when you buy an
Apple, you are paying for a concept as much as for a computer. The
Macbooks with their one-piece aluminium cases have an undeniably (in
my view) much more "solid" and robust feel to them than any other
(predominantly plastic) laptop on the market. It is definitely a
feeling of buying something designed by F.A. Porsche as opposed to
something more "mundane", and Apple have marketed it masterfully. I
personally can't think of a single laptop that can best a Macbook on
the aesthetic front.
As for desktops, it should be noted of course that the one-piece iMac
is a very different concept than a standard case-monitor desktop. The
iMac sells on very much the same principles as the Macbook: solid,
minimalistic, beautiful design, reduced clutter, and great user
experience. The average end user will probably prioritize that over
perfomance.
Geeks like me want the figures rather than the looks...but I sure
wouĺdn't mind my desktops case to be carved out of a single slate of
aluminum.
--
Best,
Christopher Svanefalk
4351
days inactive
4434
days old
82 comments
27 participants
participants (27)
-
Alan Cox -
Anthony R Fletcher -
Cameron Simpson -
Chris Adams -
Christopher Svanefalk -
Clemens Eisserer -
Darryl L. Pierce -
Dave Ihnat -
Geoffrey Leach -
Greg Woods -
Heinz Diehl -
James Wilkinson -
Jari Fredriksson -
Joe Zeff -
Joel Rees -
Marko Vojinovic -
Michael Cronenworth -
Pedro Francisco -
Rahul Sundaram -
Reindl Harald -
Rex Dieter -
Rick Stevens -
SternData -
Steve Underwood -
Tim -
Timothy Murphy -
Zind