Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.mac.system > #56676 > unrolled thread

Fan control on Xserve

Back to article view | Back to comp.sys.mac.system

Contents

  Fan control on Xserve JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-02-18 16:58 -0500
    Re: Fan control on Xserve Jolly Roger <jollyroger@pobox.com> - 2014-02-18 22:10 +0000
      Re: Fan control on Xserve JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-02-18 17:46 -0500
        Re: Fan control on Xserve billy@MIX.COM - 2014-02-19 02:37 +0000
        Re: Fan control on Xserve Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-02-20 15:18 +0000
          Re: Fan control on Xserve JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-02-20 17:10 -0500
            Re: Fan control on Xserve Lewis <g.kreme@gmail.com.dontsendmecopies> - 2014-02-21 01:37 +0000
    Re: Fan control on Xserve nospam <nospam@nospam.invalid> - 2014-02-18 17:58 -0500
      Re: Fan control on Xserve JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-02-19 00:05 -0500
        Re: Fan control on Xserve Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-02-19 17:32 -0500
          Re: Fan control on Xserve JF Mezei <jfmezei.spamnot@vaxination.ca> - 2014-02-20 01:33 -0500
            Re: Fan control on Xserve Alan Browne <alan.browne@FreelunchVideotron.ca> - 2014-02-20 17:33 -0500

#56676 — Fan control on Xserve

Does anyone know of an app that can control fans on an Xserve.

I tried smcfancontrol and it crashes/fails on the Xserve.

Ideally, I would like to test a lower default fan speed when idle,
especially in winter due to colder room.

My Xserve is only 1 CPU, so generates far less heat than what it has
been designed for and those tiny fans at 5000rpm are noisy.

Right now, I can see when fan speed starts to increase as CPU
temperature increases.  So if lowering idle fan speed a bit causes no
increase in CPU temperature I know it is safe.

Any suggestions ?

[toc] | [next] | [standalone]

#56677

On 2014-02-18, JF Mezei <jfmezei.spamnot@vaxination.ca> wrote:
> Does anyone know of an app that can control fans on an Xserve.
>
> I tried smcfancontrol and it crashes/fails on the Xserve.
>
> Ideally, I would like to test a lower default fan speed when idle,
> especially in winter due to colder room.
>
> My Xserve is only 1 CPU, so generates far less heat than what it has
> been designed for and those tiny fans at 5000rpm are noisy.
>
> Right now, I can see when fan speed starts to increase as CPU
> temperature increases.  So if lowering idle fan speed a bit causes no
> increase in CPU temperature I know it is safe.
>
> Any suggestions ?

I wouldn't advise it. I doubt they were designed with such modification
in mind. Safety should come first with expensive server-grade hardware,
which is why the fans run as fast and noisy as they do. Such hardware is
designed for data centers, where noise is the norm. I wouldn't advise
placing one in a home environment to begin with unless you are prepared
to deal with the cooling and noise associated with rack-mounted server
hardware like that. I admit I've been tempted in the past to install
rack-mounted systems in my home, but wouldn't dream of doing it without
specifically designing a closet/room for the additional cooling and
noise insulation required. For my needs, a Mac mini closet server and
Mac Pro workstation provide the server / data crunching ability quite
sufficiently for a home environment, without all the headaches of a rack
setup.

-- 
E-mail sent to this address may be devoured by my ravenous SPAM filter.
I often ignore posts from Google. Use a real news client instead.

JR

[toc] | [prev] | [next] | [standalone]

#56679

On 14-02-18 17:10, Jolly Roger wrote:

> I wouldn't advise it. I doubt they were designed with such modification
> in mind. Safety should come first with expensive server-grade hardware,

But the Xserve ws designed to have 2 CPUs. When you have only 1, the
cooling requirements are much lower, especially if the system load is
very low.

For instance, if tests would show that reducing fan speed by 1000 rpm
resulted in no temperature increase for the CPU and other areas, then it
would be perfectly safe. The rpms woudl increase normally if temperature
rises.

The xserve doesn't produce much heat. You don't put it in a closet
because that would get hit. but in a room it is fine (about 150watts of
heat, an old lightbulb).

[toc] | [prev] | [next] | [standalone]

#56691

JF Mezei <jfmezei.spamnot@vaxination.ca> writes, quoting Jolly Roger:

> > I wouldn't advise it. I doubt they were designed with such modification
> > in mind. Safety should come first with expensive server-grade hardware
> 
> But the Xserve ws designed to have 2 CPUs. When you have only 1, the
> cooling requirements are much lower, especially if the system load is
> very low.

I'd replace the fan, in this case with something less noisy.  There is
no shortage of fans from which to choose.

In my case, I went with fans that can move the most air, because the
systems are in a separate room.  Heat and dirt are the two enemies of
electronic equipment.

I also use http://www.bresink.com/osx/TemperatureMonitor.html and a
script to shutdown the systems if they overheat, as they are powered
up continuously.

Billy Y..
-- 
        sub     #'9+1   ,r0             ; convert ascii byte
	add     #9.+1   ,r0             ; to an integer
	bcc     20$                     ; not a number

[toc] | [prev] | [next] | [standalone]

#56750

In message <5303e2cb$0$10864$c3e8da3$1cbc7475@news.astraweb.com> 
  JF Mezei <jfmezei.spamnot@vaxination.ca> wrote:
> On 14-02-18 17:10, Jolly Roger wrote:

>> I wouldn't advise it. I doubt they were designed with such modification
>> in mind. Safety should come first with expensive server-grade hardware,

> But the Xserve ws designed to have 2 CPUs. When you have only 1, the
> cooling requirements are much lower, especially if the system load is
> very low.

The fan turns at the speed it needs to turn for the temperature it sees.
If the fan is too noisy, replace it with a quieter one. The Xserve was
not designed to be quite, so no money was spent on quieter fans.

> The xserve doesn't produce much heat. You don't put it in a closet
> because that would get hit. but in a room it is fine (about 150watts of
> heat, an old lightbulb).

The total heat output isn't relevant, it's the heat of the CPU.

-- 
Of course, there were various groups seeking his overthrow, and this was
right and proper and the sign of a vigorous and healthy society. No-one
could call him unreasonable about the matter. Why, hadn't he founded
most of them himself? And what was so beautiful was the way they spent
nearly all their time bickering with one another.  Human nature, the
Patrician always said, was a marvellous thing. Once you understood where
its levers were. --Guards! Guards!

[toc] | [prev] | [next] | [standalone]

#56771

On 14-02-20 10:18, Lewis wrote:

> The total heat output isn't relevant, it's the heat of the CPU.

Half the fans are aimed at empty space where the second CPU would be.

[toc] | [prev] | [next] | [standalone]

#56787

In message <53067d50$0$61271$c3e8da3$5e5e430d@news.astraweb.com> 
  JF Mezei <jfmezei.spamnot@vaxination.ca> wrote:
> On 14-02-20 10:18, Lewis wrote:

>> The total heat output isn't relevant, it's the heat of the CPU.

> Half the fans are aimed at empty space where the second CPU would be.

Are you thinking you can turn off half the fans? I don't think that is
the case.

You could slow down BOTH fans, but that is probably a bad idea.

-- 
“They always say time changes things, but you actually have to change
them yourself." Andy Warhol

[toc] | [prev] | [next] | [standalone]

#56680

In article <5303d792$0$46364$c3e8da3$a9097924@news.astraweb.com>, JF
Mezei <jfmezei.spamnot@vaxination.ca> wrote:

> Does anyone know of an app that can control fans on an Xserve.
> 
> I tried smcfancontrol and it crashes/fails on the Xserve.

try fan control, a different product, despite the similar name. it
supports up to 10.6, but that's probably not an issue with the xserve.
it says macbooks, but it works on other macs too. it includes source so
you may be able to fix whatever doesn't work.
<http://www.lobotomo.com/products/FanControl/>

or try macs fan control, yet another tool with almost the same name.
apparently coming up something original is difficult. 
<http://www.crystalidea.com/macs-fan-control>

[toc] | [prev] | [next] | [standalone]

#56704

On 14-02-18 17:58, nospam wrote:

> or try macs fan control, yet another tool with almost the same name.
> apparently coming up something original is difficult. 
> <http://www.crystalidea.com/macs-fan-control>


This one does run on the Xserve. (It can control up to 18 fans !)

While I am not able to lower the minimum speed, I was able to exercise
each fan individually.

What is interesting is that the above app shows a current RPM of 1500
for one fan, and Server Monitor shows 6.016 RPM. Roughly a factor of 4
in the values.

Running Handbrake or some Adobe encoders is about the only things that
normally cause the fans to speed up from idle, even in summer.

What I might try is to set different values for each fan to see if it
makes for quieter environment (fewer harmonics).

[toc] | [prev] | [next] | [standalone]

#56723

On 2014.02.19, 00:05 , JF Mezei wrote:
> On 14-02-18 17:58, nospam wrote:
>
>> or try macs fan control, yet another tool with almost the same name.
>> apparently coming up something original is difficult.
>> <http://www.crystalidea.com/macs-fan-control>
>
>
> This one does run on the Xserve. (It can control up to 18 fans !)
>
> While I am not able to lower the minimum speed, I was able to exercise
> each fan individually.
>
> What is interesting is that the above app shows a current RPM of 1500
> for one fan, and Server Monitor shows 6.016 RPM. Roughly a factor of 4
> in the values.

Looks to me like some fans have different methods for sensing / 
measuring speed and the factor is 4.  This may be important to how you 
control your fans.  Perhaps by listening you can figure out how to get a 
desired setting irrespective of the indicated setting.

eg: if it says "6000" and you ask for 6000 there should be no change in 
fan noise.  But it could be you need to ask for 1500 to get 6000 
indicated (which still might be 1500 - but indicating 6000).

> Running Handbrake or some Adobe encoders is about the only things that
> normally cause the fans to speed up from idle, even in summer.
>
> What I might try is to set different values for each fan to see if it
> makes for quieter environment (fewer harmonics).

Do you have something like iStat Pro running that shows you temperatures?

I'd decide on max temperatures for various parts first - then set fan 
speeds that keep those parts under that number by some margin.  Note 
that Google made a study some years ago showing that hotter hard disks 
last longer than cooler hard disks (look it up for the details).

If you just have enclosure fans, then it's better to keep the 
electronics as cool as possible.  Heat kills electronics over time and 
they're more expensive than disks.  (Did you find a way to put ordinary 
SATA disks in the Apple disk caddies?).

The real problem with your approach is that you're now serving the 
server having to watch temperatures and set fan speeds and not the other 
way around.

If noise is the issue, then putting the server in an appropriate 
location with ample ventilation - but good sound deadening would be a 
far better approach.

As to the fewer harmonics one could write a program to listen (audio 
input) to the environment, do FFT's to determine beat frequencies and 
then tune the fan RPM's to reduce the beat.  Good science project.

-- 
    ... it may be that "in the cloud" really isn't the best term
    for the services these companies offer.  What they really
    want is to have us "on the leash."
                   -David Pogue, Scientific American, 2014.02

[toc] | [prev] | [next] | [standalone]

#56740

On 14-02-19 17:32, Alan Browne wrote:

> The real problem with your approach is that you're now serving the 
> server having to watch temperatures and set fan speeds and not the other 
> way around.

Mot fan control programs on Macs are able to set the lowest fan speed to
a value equal or higher than the hardware's specified limit, but it
leaves the SMC to raise fan speed on demand.

So you only change how low the fans can go, without removing the
automation to control/raise fan speeds.

[toc] | [prev] | [next] | [standalone]

#56774

On 2014.02.20, 01:33 , JF Mezei wrote:
> On 14-02-19 17:32, Alan Browne wrote:
>
>> The real problem with your approach is that you're now serving the
>> server having to watch temperatures and set fan speeds and not the other
>> way around.
>
> Mot fan control programs on Macs are able to set the lowest fan speed to
> a value equal or higher than the hardware's specified limit, but it
> leaves the SMC to raise fan speed on demand.
>
> So you only change how low the fans can go, without removing the
> automation to control/raise fan speeds.

Could be, and that's a good compromise.  But when I was using SMC Fan 
Control it was not changing anything at any time.  Of course I was using 
it to increase speed to keep the temp of a failing HD down at the time - 
it worked fine below a given temp, but malfunctioned above that temp.

-- 
   Those who have reduced our privacy, whether they are state
   or commercial actors, prefer that we do not reduce theirs.
                  - Jaron Lanier, Scientific American, 2013.11.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.mac.system

csiph-web