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.
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
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...)
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
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.
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
On Wed, Mar 7, 2012 at 12:07 AM, Marko Vojinovic vvmarko@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@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
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@alldigital.com - - AIM/Skype: therps2 ICQ: 22643734 Yahoo: origrps2 - - - - Squawk! Pieces of Seven! Pieces of Seven! Parity Error! - ----------------------------------------------------------------------
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.
On Tue, Mar 6, 2012 at 10:13 PM, Geoffrey Leach geoff@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.
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.
Where do we get these recruits?
On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald h.reindl@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
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
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.
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
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).
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@dminet.com
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
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.
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
Once upon a time, Steve Underwood steveu@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.
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
Once upon a time, Reindl Harald h.reindl@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.
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.
Once upon a time, Steven Stern subscribed-lists@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).
Am 17.05.2012 18:33, schrieb Chris Adams:
Once upon a time, Reindl Harald h.reindl@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
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.
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
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.
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
On Thu, 17 May 2012 22:09:36 +0200 Reindl Harald h.reindl@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
Am 17.05.2012 22:19, schrieb Alan Cox:
On Thu, 17 May 2012 22:09:36 +0200 Reindl Harald h.reindl@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
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.
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
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.
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
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.
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@dminet.com
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
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.
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
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.
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.
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. :)
On Thu, May 17, 2012 at 3:24 PM, Michael Cronenworth mike@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@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
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.
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
On 17May2012 08:30, Dave Ihnat dihnat@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.
On 17May2012 23:52, Steve Underwood steveu@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.
On 17May2012 11:33, Chris Adams cmadams@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,
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.
On 17May2012 22:33, Reindl Harald h.reindl@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.
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..
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.
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
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
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
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
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").
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
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.
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?
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.
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.
On Sat, May 19, 2012 at 9:29 AM, Heinz Diehl htd@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@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
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@dminet.com
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)
On Sat, May 19, 2012 at 2:30 PM, Reindl Harald h.reindl@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@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.
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.
On Sat, May 19, 2012 at 8:41 PM, Joe Zeff joe@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@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
On Thu, May 17, 2012 at 9:36 PM, Reindl Harald h.reindl@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
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
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)
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
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@gmail.com Where do we get these recruits? On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald h.reindl@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@gmail.com
please guys, start your private flamefest and leave us alone
2012/5/26 Reindl Harald h.reindl@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@gmail.com Where do we get these recruits? On Wed, Mar 7, 2012 at 7:35 AM, Reindl Harald h.reindl@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@gmail.com
-- users mailing list users@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
On Sat, May 26, 2012 at 5:20 PM, Reindl Harald h.reindl@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
Stop this! (please)
2012/5/28 Joel Rees joel.rees@gmail.com:
On Sat, May 26, 2012 at 5:20 PM, Reindl Harald h.reindl@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@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