-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Only for info...;o) today I had time to try with kernel2.4.20-rc1
(linux-2.4.19.tar.gz + patch-2.4.20-rc1.gz)... I take .configure ~ from
kernel-enterprise-2.4.19.18mdk-1-1mdk....compiled and installed.....and now I have
correct behaviour !!!!!....;o)
NO MORE HIGH SYSTEM CPU (and little more speed...;o) after more than a week with first
sys-cpu always at more than 90% doing nothing....
cu....and good work....or may be.... good weekend....(all togheter is
better!!!!).......;o)
- ----------------------------
Linux Version 2.4.20-rc1HalfCat, Compiled #2 SMP Fri Nov 15 19:32:10 CET 2002
Two 1.98GHz Intel Intel(R) Xeon(TM) CPU 2.00GHz Processors, 1024M RAM, 7890.52
Bogomips Total
- ---
19:59:01 up 4 min, 25 users, load average: 1.40, 0.70, 0.27
105 processes: 103 sleeping, 2 running, 0 zombie, 0 stopped
CPU0: 0.1% user 0.0% system 0.0% nice 0.0% iowait 99.8% idle <===NO MORE
HIGH SYSTEM CPU
CPU1: 0.0% user 0.3% system 0.0% nice 0.0% iowait 99.6% idle ===doing
nothing
Mem: 1033096k av, 241444k used, 791652k free, 0k shrd, 9188k buff
50972k active, 157088k inactive
Swap: 1052216k av, 0k used, 1052216k free 104940k cached
-
---------------------------------------------------------------------------------------------------------------------------------------------
On Wednesday 06 November 2002 09:49, Franco Silvestro wrote:
> I wrote that because I noticed for days (from time I installed -18
> kernel...) that first cpu (only first...?... and with gkrellm is much
> evident) stay always on high system-cpu on my dual xeon machine( compaq evo
> w8000, the only test smp I've) ....also when I really do nothing (and in
> that time I expect cpus about 99% idle...) ...I thinked about some kind
> of bad thread management in smp environment with new modules/new kernel
> version (on my single cpu system laptop I haven't that behaviour with
> -18...doing nothing, cpu stay normally on 99% idle)
>
> Now I've tried compile(nessus2 and nessus3 at the same time) on -16
> enterprise and see below.... I've 100% used on both cpus....and after
> compile ending both cpus return at ~99% idle it seems correct.....no? (I
> haven't noticed on my previous email that 0% idle in top output.....really
> strange !?!):
>
> during compile:
> 08:31:23 up 20:09, 24 users, load average: 0.43, 0.21, 0.34
> 103 processes: 101 sleeping, 2 running, 0 zombie, 0 stopped
> 09:21:06 up 20:59, 24 users, load average: 0.71, 0.27, 0.13
> 114 processes: 108 sleeping, 5 running, 1 zombie, 0 stopped
> CPU0: 76.5% user 23.6% system 0.0% nice 0.0% iowait 0.0% idle
> CPU1: 74.9% user 25.0% system 0.0% nice 0.0% iowait 0.0% idle
> Mem: 1033364k av, 1009132k used, 24232k free, 0k shrd, 33892k
> buff 280652k active, 615408k inactive
> Swap: 1052216k av, 4064k used, 1048152k free 744316k
> cached
>
> after compile:
> 08:31:23 up 20:09, 24 users, load average: 0.43, 0.21, 0.34
> 103 processes: 101 sleeping, 2 running, 0 zombie, 0 stopped
> 09:21:06 up 20:59, 24 users, load average: 0.71, 0.27, 0.13
> 114 processes: 108 sleeping, 5 running, 1 zombie, 0 stopped
> 09:37:48 up 21:16, 24 users, load average: 0.00, 0.02, 0.06
> 102 processes: 101 sleeping, 1 running, 0 zombie, 0 stopped
> CPU0: 0.3% user 0.3% system 0.0% nice 0.0% iowait 99.2% idle
> CPU1: 0.7% user 0.5% system 0.0% nice 0.0% iowait 98.6% idle
> Mem: 1033364k av, 1005420k used, 27944k free, 0k shrd, 33836k
> buff 285716k active, 606620k inactive
> Swap: 1052216k av, 4064k used, 1048152k free 744048k
> cached ...cu....and good work....;o)
>
> On Tuesday 05 November 2002 20:57, Todd Lyons wrote:
> > Franco Silvestro wrote on Tue, Nov 05, 2002 at 01:24:39PM +0100 :
> > > I see that with last kernel-enterprise (doing nothing), very high
> > > system cpu: CPU0: 0.5% user 84.4% system 0.0% nice 0.0% iowait
> > > 14.9% idle CPU1: 9.5% user 10.9% system 0.0% nice 0.0% iowait
> > > 79.4% idle
[...]
> > You've highlighted something I saw on another machine. Notice how on
> > the .16mdk the numbers don't add up? Even in a heavy compile, the
> > numbers won't add up to 100%. There seems to be some kind of scaling
> > that is happening.
> >
> > Blue skies... Todd
- --
Franco Silvestro - Servizio di Sicurezza reti e dati
c/o CeSIA - Alma Mater Studiorum - Universita' di Bologna
'lynx -source http://www.reti.unibo.it/FrancoSilvestro.asc | gpg --import'
Key fingerprint = 3145 1309 C5D8 7B3F 2582 44D2 91BB ACAB E010 ABF9
Current Linux uptime: 0 hours 6 minutes.
- -----------------------------------------
Fri Nov 15 20:01:04 CET 2002
- -----------------------------------------
Linux 2.4.20-rc1HalfCat GNU/Linux
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
iD8DBQE91Ul+kbusq+AQq/kRAn5sAJ47lGM88n4q/8kUkXXR8Rb1kCt7RgCfV3v/
let6a4hwUqTMC/qwTIqyCvI=
=3GtS
-----END PGP SIGNATURE-----
