Gmail Calendar Documents Reader Web more »
Recently Visited Groups | Help | Sign in
Google Groups Home
"uptime - sto je load average"
There are currently too many topics in this group that display first. To make this topic appear first, remove this option from another topic.
There was an error processing your request. Please try again.
flag
  Messages 1 - 25 of 55 - Collapse all  -  Translate all to Translated (View all originals)   Newer >
The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
 
From:
To:
Cc:
Followup To:
Add Cc | Add Followup-to | Edit Subject
Subject:
Validation:
For verification purposes please type the characters you see in the picture below or the numbers you hear by clicking the accessibility icon. Listen and type the numbers you hear
 
Valent Turkovic  
View profile  
 More options Oct 6 2008, 8:36 am
Newsgroups: hr.comp.os.linux
From: Valent Turkovic <valent.turko...@gmail.com>
Date: Mon, 6 Oct 2008 12:36:37 +0000 (UTC)
Local: Mon, Oct 6 2008 8:36 am
Subject: "uptime - sto je load average"
http://mzsl.wordpress.com/2008/09/24/uptime-sto-je-load-average/

cini mi se da je clanak malo fulao jer citajuci:
http://www.teamquest.com/resources/gunther/display/5/
pise explicitno:

You would be forgiven for jumping to the conclusion that the "load" is
the same thing as the CPU utilization. As the Linux results show, when
two hot processes are running, the maximum load is two (not one) on a
single CPU. So, load is not equivalent to CPU utilization.

Valent.

--
http://kernelreloaded.blog385.com/
linux, blog, anime, spirituality, windsurf, wireless
registered as user #367004 with the Linux Counter, http://counter.li.org.
ICQ: 2125241, Skype: valent.turkovic


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Valent Turkovic  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 8:54 am
Newsgroups: hr.comp.os.linux
From: Valent Turkovic <valent.turko...@gmail.com>
Date: Mon, 6 Oct 2008 12:54:39 +0000 (UTC)
Local: Mon, Oct 6 2008 8:54 am
Subject: Re: "uptime - sto je load average"

On Mon, 06 Oct 2008 12:36:37 +0000, Valent Turkovic wrote:
> cini mi se da je clanak malo fulao jer citajuci:

Eh, ja sam prebrzo iscitao i krivo zakljucio :)

Ono sto me zbunjuje je to da imam veliki load average na stroju koji je
vrlo malo opterecen (4 aktivna torrenta), opterecenje procesora je 20-30%
ali load je vrlo velik - 3.59   4.12   3.90

Procesor je AMD Barton 2500+ (radi na 2000Mhz) i sistem ima 512MB
memorije.

Dok na drugom stroju koji je jako opterečen (firefox s 20-30 tabova,
virtualna mašiona, još jedno 5-6 traznih aplikacija, torrenti...) s 2GB
rama i kojem je CPU stalno na 90-100% load je  - 2.12,   2.53,   2.75

Valent.

--
http://kernelreloaded.blog385.com/
linux, blog, anime, spirituality, windsurf, wireless
registered as user #367004 with the Linux Counter, http://counter.li.org.
ICQ: 2125241, Skype: valent.turkovic


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Vedran Furač  
View profile  
 More options Oct 6 2008, 8:58 am
Newsgroups: hr.comp.os.linux
From: Vedran Furač <vedr...@vedranf.mine.nu>
Date: Mon, 06 Oct 2008 14:58:30 +0200
Local: Mon, Oct 6 2008 8:58 am
Subject: Re: "uptime - sto je load average"

Valent Turkovic wrote:
> http://mzsl.wordpress.com/2008/09/24/uptime-sto-je-load-average/

> cini mi se da je clanak malo fulao jer citajuci:
> http://www.teamquest.com/resources/gunther/display/5/
> pise explicitno:

> You would be forgiven for jumping to the conclusion that the "load" is
> the same thing as the CPU utilization. As the Linux results show, when
> two hot processes are running, the maximum load is two (not one) on a
> single CPU. So, load is not equivalent to CPU utilization.

Što je i jasno. Čovjek koji piše blog je mogao napraviti jednostavan
test: kopirati fajl s jedne particije na drugu. Procesor će biti
minimalno opterećen, a LA će skočiti. Kod mene, na kraju kopiranja 3GB
fajla s jedne ext3 particije na drugu LA je skočio na 10 (cfq io sched).

    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Matija Papec  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 9:17 am
Newsgroups: hr.comp.os.linux
From: Matija Papec <n...@yahoo.com>
Date: Mon, 06 Oct 2008 15:17:55 +0200
Local: Mon, Oct 6 2008 9:17 am
Subject: Re: "uptime - sto je load average"

Valent Turkovic wrote:
>> cini mi se da je clanak malo fulao jer citajuci:

> Eh, ja sam prebrzo iscitao i krivo zakljucio :)

> Ono sto me zbunjuje je to da imam veliki load average na stroju koji je
> vrlo malo opterecen (4 aktivna torrenta), opterecenje procesora je 20-30%
> ali load je vrlo velik - 3.59   4.12   3.90

> Procesor je AMD Barton 2500+ (radi na 2000Mhz) i sistem ima 512MB
> memorije.

To je i mene cudilo na jednoj masini, ali se ispostavilo da na momente
disk nesto intezivno vrti dok cpu idla (ili odlazi u wait).

    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Nikola Ostrun  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 9:27 am
Newsgroups: hr.comp.os.linux
From: Nikola Ostrun <nostrun-use...@foi.hr>
Date: Mon, 6 Oct 2008 13:27:57 +0000 (UTC)
Local: Mon, Oct 6 2008 9:27 am
Subject: Re: "uptime - sto je load average"
Dana 2008-10-06, Valent Turkovic <valent.turko...@gmail.com> je napisao:

Pogledaj sa sar(1)-om na sto tocno se 'trosi' cpu.

    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Dinko Korunic  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 9:31 am
Newsgroups: hr.comp.os.linux
From: Dinko Korunic <dinko.koru...@gmail.com>
Date: Mon, 6 Oct 2008 15:31:40 +0200
Local: Mon, Oct 6 2008 9:31 am
Subject: Re: "uptime - sto je load average"

On Mon, 06 Oct 2008 14:58:30 +0200, Vedran Furač wrote:
> Što je i jasno. Čovjek koji piše blog je mogao napraviti jednostavan

Jasno je bilo kome tko je ucio osnovne arhitekture operacijskih sustava,
ali ocito covjek to nije citao..

> test: kopirati fajl s jedne particije na drugu. Procesor će biti
> minimalno opterećen, a LA će skočiti. Kod mene, na kraju kopiranja 3GB

Nece biti minimalno optecen, osim ako si mislio samo na user-time?

> fajla s jedne ext3 particije na drugu LA je skočio na 10 (cfq io sched).

Kernel time se povecao, broj slotova za user-time se smanjio, broj
procesa/taskova u redu za cekanje se povecao.

--
NAME:Dinko.kreator.Korunic   DISCLAIMER:Standard.disclaimer.applies
ICQ:16965294        JAB:kreator...@jabber.org        PGP:0xEA160D0B
HOME:http://dkorunic.net    QUOTE:Eat.right.stay.fit.and.die.anyway


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Dinko Korunic  
View profile  
 More options Oct 6 2008, 9:32 am
Newsgroups: hr.comp.os.linux
From: Dinko Korunic <dinko.koru...@gmail.com>
Date: Mon, 6 Oct 2008 15:32:43 +0200
Local: Mon, Oct 6 2008 9:32 am
Subject: Re: "uptime - sto je load average"

On Mon, 6 Oct 2008 13:27:57 +0000 (UTC), Nikola Ostrun wrote:
> Pogledaj sa sar(1)-om na sto tocno se 'trosi' cpu.

Definitivno bi vise saznao sa atop-om, sar je so 1980s :-)

--
NAME:Dinko.kreator.Korunic   DISCLAIMER:Standard.disclaimer.applies
ICQ:16965294        JAB:kreator...@jabber.org        PGP:0xEA160D0B
HOME:http://dkorunic.net    QUOTE:Eat.right.stay.fit.and.die.anyway


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Zeba  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 12:00 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Zeba <hrvoje.zebaMA...@MEmail.chihq.org>
Date: Mon, 06 Oct 2008 18:00:29 +0200
Local: Mon, Oct 6 2008 12:00 pm
Subject: Re: "uptime - sto je load average"

Dinko Korunic wrote:
> /.../

khm... a sta nije load prosjecan broj aktivnih threadova u redu za
izvrsavanje?

--
I doubt, therefore I might be.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Igor Pozgaj  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 12:19 pm
Newsgroups: hr.comp.os.linux
From: Igor Pozgaj <ipoz...@nospam.fly.srk.fer.hr>
Date: Mon, 6 Oct 2008 16:19:00 +0000 (UTC)
Subject: Re: "uptime - sto je load average"

On Mon, 06 Oct 2008 18:00:29 +0200, Hrvoje Zeba wrote:
> Dinko Korunic wrote:
>> /.../

> khm... a sta nije load prosjecan broj aktivnih threadova u redu za
> izvrsavanje?

Je, ali taj broj ti nece puno toga reci sam po sebi. Recimo ja sam
nedavno na poslu imao situaciju da je stroj (AIX, 24 corea, 128 GB
memorije, FC diskovi) imao load average oko 270, a radio je cisto
ok. Stvari koje ljudi zaboravljaju su da je to *average* (usporedi
sa prvom kolonom u ispisu naredbe vmstat 60) te da se interpretacija
broja mijenja s vecim brojem kora/procesora.

--
Igor Pozgaj | ipozgaj at fly.srk.fer.hr
ICQ: 126002505  | IRC: @thunder (#linux@IdolNet)
PGP: 0xEF36A092 | http://fly.srk.fer.hr/~ipozgaj
http://ipozgaj.blogspot.com (/atom.xml RSS feed)


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Zeba  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 12:45 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Zeba <hrvoje.zebaMA...@MEmail.chihq.org>
Date: Mon, 06 Oct 2008 18:45:55 +0200
Local: Mon, Oct 6 2008 12:45 pm
Subject: Re: "uptime - sto je load average"

Igor Pozgaj wrote:
> On Mon, 06 Oct 2008 18:00:29 +0200, Hrvoje Zeba wrote:

>> Dinko Korunic wrote:
>>> /.../
>> khm... a sta nije load prosjecan broj aktivnih threadova u redu za
>> izvrsavanje?

> Je, ali taj broj ti nece puno toga reci sam po sebi. Recimo ja sam
> nedavno na poslu imao situaciju da je stroj (AIX, 24 corea, 128 GB
> memorije, FC diskovi) imao load average oko 270, a radio je cisto
> ok. Stvari koje ljudi zaboravljaju su da je to *average* (usporedi
> sa prvom kolonom u ispisu naredbe vmstat 60) te da se interpretacija
> broja mijenja s vecim brojem kora/procesora.

to ti je cca 11 threadova po procesoru sto ne bi trebalo biti strasno (a
i ovisi o vrsti threadova koji se vrte) a trebao bi i pogledati sto se
sve na AIX-u smatra aktivnim threadom.

--
I doubt, therefore I might be.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jakov Sosic  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 1:53 pm
Newsgroups: hr.comp.os.linux
From: Jakov Sosic <jso...@jsosic.homeunix.org>
Date: Mon, 6 Oct 2008 19:53:21 +0200
Local: Mon, Oct 6 2008 1:53 pm
Subject: Re: "uptime - sto je load average"
On 2008-10-06, Igor Pozgaj <ipoz...@nospam.fly.srk.fer.hr> wrote:

> Je, ali taj broj ti nece puno toga reci sam po sebi. Recimo ja sam
> nedavno na poslu imao situaciju da je stroj (AIX, 24 corea, 128 GB
> memorije, FC diskovi) imao load average oko 270, a radio je cisto
> ok. Stvari koje ljudi zaboravljaju su da je to *average* (usporedi
> sa prvom kolonom u ispisu naredbe vmstat 60) te da se interpretacija
> broja mijenja s vecim brojem kora/procesora.

Pa naravno da stroj radi ok - kad je moguce da 260 od tih 270 ceka
na "izlazak sunca" i time samo stoji u queue i uopce ne opterecuju
stroj... Load zna strasno zavarati...

--
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Nikola Ostrun  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 2:19 pm
Newsgroups: hr.comp.os.linux
From: nikola.ost...@email.t-com.hr (Nikola Ostrun)
Date: Mon, 06 Oct 2008 20:19:24 +0200
Local: Mon, Oct 6 2008 2:19 pm
Subject: Re: "uptime - sto je load average"

Dinko Korunic <dinko.koru...@gmail.com> writes:
> On Mon, 6 Oct 2008 13:27:57 +0000 (UTC), Nikola Ostrun wrote:
>> Pogledaj sa sar(1)-om na sto tocno se 'trosi' cpu.

> Definitivno bi vise saznao sa atop-om, sar je so 1980s :-)

  Zašto? Da vidi postotke CPU usagea po kategorijama mu je sar(1)
savršen. Doduše, nisam mislio da je pregled po procesu bitan...

--
VI = Virtually Incomprehensible.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Nikola Ostrun  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 2:21 pm
Newsgroups: hr.comp.os.linux
From: nikola.ost...@email.t-com.hr (Nikola Ostrun)
Date: Mon, 06 Oct 2008 20:21:36 +0200
Local: Mon, Oct 6 2008 2:21 pm
Subject: Re: "uptime - sto je load average"

Jakov Sosic <jso...@jsosic.homeunix.org> writes:
> On 2008-10-06, Igor Pozgaj <ipoz...@nospam.fly.srk.fer.hr> wrote:

>> Je, ali taj broj ti nece puno toga reci sam po sebi. Recimo ja sam
>> nedavno na poslu imao situaciju da je stroj (AIX, 24 corea, 128 GB
>> memorije, FC diskovi) imao load average oko 270, a radio je cisto
>> ok. Stvari koje ljudi zaboravljaju su da je to *average* (usporedi
>> sa prvom kolonom u ispisu naredbe vmstat 60) te da se interpretacija
>> broja mijenja s vecim brojem kora/procesora.

> Pa naravno da stroj radi ok - kad je moguce da 260 od tih 270 ceka
> na "izlazak sunca" i time samo stoji u queue i uopce ne opterecuju
> stroj... Load zna strasno zavarati...

  Ako proces/dretva *čeka* nešto (ako joj stanje nije RUNNABLE ili
kak se već to zove na linuxu) onda nije /load/ :-)

--
VI = Virtually Incomprehensible.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Vedran Furač  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 2:48 pm
Newsgroups: hr.comp.os.linux
From: Vedran Furač <vedr...@vedranf.mine.nu>
Date: Mon, 06 Oct 2008 20:48:15 +0200
Local: Mon, Oct 6 2008 2:48 pm
Subject: Re: "uptime - sto je load average"

Ali ulazi u izračun LA na linuxu. Zato se i dobivaju ove "čudne" i
visoke brojke.

    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Zeba  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 2:53 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Zeba <hrvoje.zebaMA...@MEmail.chihq.org>
Date: Mon, 06 Oct 2008 20:53:49 +0200
Local: Mon, Oct 6 2008 2:53 pm
Subject: Re: "uptime - sto je load average"

Vedran Furač wrote:
>>>> Je, ali taj broj ti nece puno toga reci sam po sebi. Recimo ja sam
>>>> nedavno na poslu imao situaciju da je stroj (AIX, 24 corea, 128 GB
>>>> memorije, FC diskovi) imao load average oko 270, a radio je cisto
>>>> ok. Stvari koje ljudi zaboravljaju su da je to *average* (usporedi
>>>> sa prvom kolonom u ispisu naredbe vmstat 60) te da se interpretacija
>>>> broja mijenja s vecim brojem kora/procesora.
>>> Pa naravno da stroj radi ok - kad je moguce da 260 od tih 270 ceka
>>> na "izlazak sunca" i time samo stoji u queue i uopce ne opterecuju
>>> stroj... Load zna strasno zavarati...
>>   Ako proces/dretva *čeka* nešto (ako joj stanje nije RUNNABLE ili
>> kak se već to zove na linuxu) onda nije /load/ :-)

> Ali ulazi u izračun LA na linuxu. Zato se i dobivaju ove "čudne" i
> visoke brojke.

pozgaj je govorio za aix ;)

--
I doubt, therefore I might be.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Vedran Furač  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 3:08 pm
Newsgroups: hr.comp.os.linux
From: Vedran Furač <vedr...@vedranf.mine.nu>
Date: Mon, 06 Oct 2008 21:08:00 +0200
Subject: Re: "uptime - sto je load average"

OK, ne znam ništa o AIX-u. Ovo "čudne i visoke brojke" se odnosilo na
onih mojih 10 kad kopiram fajl.

    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Vedran Furač  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 3:11 pm
Newsgroups: hr.comp.os.linux
From: Vedran Furač <vedr...@vedranf.mine.nu>
Date: Mon, 06 Oct 2008 21:11:56 +0200
Local: Mon, Oct 6 2008 3:11 pm
Subject: Re: "uptime - sto je load average"

Dinko Korunic wrote:
> On Mon, 06 Oct 2008 14:58:30 +0200, Vedran Furač wrote:
>> Što je i jasno. Čovjek koji piše blog je mogao napraviti jednostavan

> Jasno je bilo kome tko je ucio osnovne arhitekture operacijskih sustava,
> ali ocito covjek to nije citao..

>> test: kopirati fajl s jedne particije na drugu. Procesor će biti
>> minimalno opterećen, a LA će skočiti. Kod mene, na kraju kopiranja 3GB

> Nece biti minimalno optecen, osim ako si mislio samo na user-time?

Ovisi o tome što je to "minimalno". U mom slučaju je to ~10%. To bi
trebalo bi ostaviti dosta prostora za ostale procese.

>> fajla s jedne ext3 particije na drugu LA je skočio na 10 (cfq io sched).

> Kernel time se povecao, broj slotova za user-time se smanjio, broj
> procesa/taskova u redu za cekanje se povecao.

Da.

Btw., proces cp je većinu vremena u "uninterruptible sleep" stanju (D u
ps-u). Možda i više nego što bi trebao pa to dodatno digne LA.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Kusulja  
View profile  
 More options Oct 6 2008, 4:48 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Kusulja <hkusuljaNOS...@kusulja.com>
Date: Mon, 6 Oct 2008 22:48:24 +0200
Local: Mon, Oct 6 2008 4:48 pm
Subject: Re: "uptime - sto je load average"

On Mon, 6 Oct 2008 12:36:37 +0000 (UTC), Valent Turkovic wrote:
> http://mzsl.wordpress.com/2008/09/24/uptime-sto-je-load-average/

> cini mi se da je clanak malo fulao jer citajuci:
> http://www.teamquest.com/resources/gunther/display/5/
> pise explicitno:

> You would be forgiven for jumping to the conclusion that the "load" is
> the same thing as the CPU utilization. As the Linux results show, when
> two hot processes are running, the maximum load is two (not one) on a
> single CPU. So, load is not equivalent to CPU utilization.

> Valent.

ok mene je sada zbunio taj blog s vasim komentarima...

imam dakle quad core, 8gb ram, 64bit ubuntu-server..

load average: 4.17, 4.75, 4.81

dal to znaci da je u prosjeku u 1min, 5 min i 15 min, moj server
preopterecen ?

ili sam nesto krivo shvatio..
tj. drugim rjecima, dal je blog tocan ili nije tocan ? (ako nije, sta je
onda tocno objasnjenje tih triju brojki)
hvala..

--
Hrvoje Kusulja
TSRB, CCNA, TVZ-Informatika


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Dinko Korunic  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 4:58 pm
Newsgroups: hr.comp.os.linux
From: Dinko Korunic <dinko.koru...@gmail.com>
Date: Mon, 6 Oct 2008 22:58:14 +0200
Local: Mon, Oct 6 2008 4:58 pm
Subject: Re: "uptime - sto je load average"

On Mon, 6 Oct 2008 22:48:24 +0200, Hrvoje Kusulja wrote:
> imam dakle quad core, 8gb ram, 64bit ubuntu-server..
> load average: 4.17, 4.75, 4.81
> dal to znaci da je u prosjeku u 1min, 5 min i 15 min, moj server
> preopterecen ?

Totalno. Trebas kupit jos jedan quadcore i prijeci na 128bitnu distru :-)

--
NAME:Dinko.kreator.Korunic   DISCLAIMER:Standard.disclaimer.applies
ICQ:16965294        JAB:kreator...@jabber.org        PGP:0xEA160D0B
HOME:http://dkorunic.net    QUOTE:Eat.right.stay.fit.and.die.anyway


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Zeba  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 5:33 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Zeba <hrvoje.zebaMA...@MEmail.chihq.org>
Date: Mon, 06 Oct 2008 23:33:33 +0200
Local: Mon, Oct 6 2008 5:33 pm
Subject: Re: "uptime - sto je load average"

Dinko Korunic wrote:
> On Mon, 6 Oct 2008 22:48:24 +0200, Hrvoje Kusulja wrote:
>> imam dakle quad core, 8gb ram, 64bit ubuntu-server..
>> load average: 4.17, 4.75, 4.81
>> dal to znaci da je u prosjeku u 1min, 5 min i 15 min, moj server
>> preopterecen ?

> Totalno. Trebas kupit jos jedan quadcore i prijeci na 128bitnu distru :-)

mudro :)

--
I doubt, therefore I might be.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Nikola Ostrun  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 5:51 pm
Newsgroups: hr.comp.os.linux
From: nikola.ost...@email.t-com.hr (Nikola Ostrun)
Date: Mon, 06 Oct 2008 23:51:48 +0200
Local: Mon, Oct 6 2008 5:51 pm
Subject: Re: "uptime - sto je load average"

Vedran Furač <vedr...@vedranf.mine.nu> writes:
> Nikola Ostrun wrote:
...
>>   Ako proces/dretva *čeka* nešto (ako joj stanje nije RUNNABLE ili
>> kak se već to zove na linuxu) onda nije /load/ :-)

> Ali ulazi u izračun LA na linuxu. Zato se i dobivaju ove "čudne" i
> visoke brojke.

  HAH, na linuxu se stvarno uzimaju. Ne sve koje čekaju, nego samo
one sa stanjem TASK_UNINTERRUPTIBLE. Gledam u kern/timer.c funkciju
calc_load() koja na kraju poziva nr_active() iz kern/sched.c da sazna
broj "aktivnih" dretvi:

unsigned long nr_active(void)
{
    unsigned long i, running = 0, uninterruptible = 0;

    for_each_online_cpu(i) {
        running += cpu_rq(i)->nr_running;
        uninterruptible += cpu_rq(i)->nr_uninterruptible;
    }

    if (unlikely((long)uninterruptible < 0))
        uninterruptible = 0;

    return running + uninterruptible;

}

  i sad me zanima što je tako posebno sa TASK_UNINTERRUPTIBLE
dretvama da ih se smatra aktivnima...

  Inače, u kern/sched.c postoji nr_running() koja je slična
nr_active() samo što ne broji TASK_UNINTERRUPTIBLE dretve. Ak te muči
visok load zamijeni u kern/timer.c u funkciji count_active_tasks()
nr_active() sa nr_running() :-)

--
VI = Virtually Incomprehensible.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Vedran Furač  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 6:09 pm
Newsgroups: hr.comp.os.linux
From: Vedran Furač <vedr...@vedranf.mine.nu>
Date: Tue, 07 Oct 2008 00:09:36 +0200
Local: Mon, Oct 6 2008 6:09 pm
Subject: Re: "uptime - sto je load average"

Nikola Ostrun wrote:
>   Inače, u kern/sched.c postoji nr_running() koja je slična
> nr_active() samo što ne broji TASK_UNINTERRUPTIBLE dretve. Ak te muči
> visok load zamijeni u kern/timer.c u funkciji count_active_tasks()
> nr_active() sa nr_running() :-)

:-)

Ne mući me to, već to što svako kopiranje veće količine podataka
značajno uspori sve ostale operacije.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Hrvoje Zeba  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 6:12 pm
Newsgroups: hr.comp.os.linux
From: Hrvoje Zeba <hrvoje.zebaMA...@MEmail.chihq.org>
Date: Tue, 07 Oct 2008 00:12:54 +0200
Local: Mon, Oct 6 2008 6:12 pm
Subject: Re: "uptime - sto je load average"

Vedran Furač wrote:
> Nikola Ostrun wrote:

>>   Inače, u kern/sched.c postoji nr_running() koja je slična
>> nr_active() samo što ne broji TASK_UNINTERRUPTIBLE dretve. Ak te muči
>> visok load zamijeni u kern/timer.c u funkciji count_active_tasks()
>> nr_active() sa nr_running() :-)

> :-)

> Ne mući me to, već to što svako kopiranje veće količine podataka
> značajno uspori sve ostale operacije.

a da te slucajno ne pati DMA?

--
I doubt, therefore I might be.


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jakov Sosic  
View profile  
 More options Oct 6 2008, 6:19 pm
Newsgroups: hr.comp.os.linux
From: Jakov Sosic <jso...@jsosic.homeunix.org>
Date: Tue, 7 Oct 2008 00:19:04 +0200
Local: Mon, Oct 6 2008 6:19 pm
Subject: Re: "uptime - sto je load average"
On 2008-10-06, Dinko Korunic <dinko.koru...@gmail.com> wrote:

> Totalno. Trebas kupit jos jedan quadcore i prijeci na 128bitnu distru :-)

Ako ne OS onda barem FS ;)

--
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Jakov Sosic  
View profile   Translate to Translated (View Original)
 More options Oct 6 2008, 6:18 pm
Newsgroups: hr.comp.os.linux
From: Jakov Sosic <jso...@jsosic.homeunix.org>
Date: Tue, 7 Oct 2008 00:18:01 +0200
Local: Mon, Oct 6 2008 6:18 pm
Subject: Re: "uptime - sto je load average"
On 2008-10-06, Nikola Ostrun <nikola.ost...@email.t-com.hr> wrote:

>   Ako proces/dretva *čeka* nešto (ako joj stanje nije RUNNABLE ili
> kak se već to zove na linuxu) onda nije /load/ :-)

Na Linuxu je... Za ostale OS-ove neznam...

--
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |


    Reply to author    Forward  
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Messages 1 - 25 of 55   Newer >
« Back to Discussions « Newer topic     Older topic »

Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
©2009 Google