Re: Interrupt using all the CPU

On Tuesday 04 July 2006 13:10, nocturnal wrote:
Hi

The -m argument exists in 5.3-RELEASE, thank you. I've attached one file
with -mio and the other without. With -mio it just output some extra
stuff at the beginning so i don't know how much it matters or maybe i
did something wrong.

I see nothing backing up those numbers...

You said it doesn't seem to affect the web server, That everything
feels normal. You cannot feel the load using the box, typing, etc, right?
In cases where interrupts reach 25%, the load is very easily noticable,
that means you have trouble typing...I would say that such a percentage
(100%) is false, you wouldn't be able to log in to the box, the kernel itself
would have trouble doing something, but interrupt handling.

A wild guess would be that you hit a bug. After all 5.3-RELEASE was the
first release marked STABLE from the 5 branch.




Med vänliga hälsningar

Stefan Midjich aka nocturnal
[Swehack] http://swehack.se

Nikos Vassiliadis wrote:
On Tuesday 04 July 2006 11:40, nocturnal wrote:
Hi

This is what i get, i'm afraid i'm not good at interpreting the results.

procs memory page disks faults
cpu r b w avm fre flt re pi po fr sr ad4 ad6 in sy cs
us sy id
1 5 0 274324 36436 26 0 0 0 50 27 0 0 242 149 332 0
88 12
1 5 0 274324 36424 3 0 0 0 3 0 1 0 338 911 462 0
50 50
1 5 0 274324 36424 0 0 0 0 0 0 16 0 391 1011 581 0
33 67
1 5 0 274324 36424 0 0 0 0 3 0 1 0 611 263 877 0
33 67
1 5 0 274324 36424 0 0 0 0 0 0 1 0 674 128 931 0
75 25
1 5 0 274324 36424 0 0 0 0 0 0 0 0 563 118 711 0
0 100
1 5 0 274324 36424 0 0 0 0 0 0 0 0 531 129 702 0
67 33
1 5 0 274324 36424 0 0 0 0 0 0 0 0 440 118 578 0
50 50
1 6 0 274324 36376 0 0 0 0 0 0 0 0 563 158 774 0
67 33
1 5 0 274324 36268 0 0 0 0 0 0 10 0 702 1019 920 0
100 0
1 5 0 274324 36268 0 0 0 0 0 0 0 0 683 118 872 0
25 75
0 6 0 291460 33800 955 0 0 0 333 0 0 0 783 1423 1069
0 50 50
0 5 0 274324 36240 611 0 0 0 1500 0 0 0 491 567 707
0 100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 717 121 986 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 729 129 923 0
33 67
0 5 0 274324 36240 0 0 0 0 0 0 0 0 570 118 742 0
50 50
0 5 0 274324 36240 0 0 0 0 0 0 0 0 568 1381 793 0
33 67
0 5 0 274324 36240 0 0 0 0 4 0 32 0 716 129 1036
0 75 25
0 5 0 274324 36240 0 0 0 0 0 0 0 0 745 118 843 0
75 25
0 5 0 274324 36240 0 0 0 0 0 0 7 0 610 137 820 0
67 33
0 5 0 274324 36240 0 0 0 0 0 0 0 0 465 174 630 0
0 100
0 5 0 274324 36240 0 0 0 0 0 0 0 0 327 237 454 0
50 50
0 5 0 274324 36240 0 0 0 0 0 0 0 0 253 125 323 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 249 127 314 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 240 118 303 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 238 125 301 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 255 129 330 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 258 144 337 0
100 0
0 5 0 274324 36240 0 0 0 0 0 0 0 0 269 129 358 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 2 0 312 527 431 0
100 0
0 5 0 274324 36236 0 0 0 0 12 0 21 0 287 129 409 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 0 0 257 128 331 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 0 0 253 118 324 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 0 0 248 128 315 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 5 0 257 118 333 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 0 0 257 118 327 0
100 0
0 5 0 274324 36236 0 0 0 0 0 0 0 0 294 118 366 0
100 0

Please turn off word wrapping on your mail program when
you send such output. Or send as attachment. Well, normal
number of interrupts, a lot of time in kernel, quite a few context
switches

send as attachment "top -n -S -I -d 10" please
There is also a -m in 6.x, don't know about 5.3-R, but it would be
interesting if it exists

from top manual:
-mdisplay
Display either cpu or io statistics. Default is cpu.
_______________________________________________
freebsd-questions@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscribe@xxxxxxxxxxx"

Relevant Pages

  • Re: [parisc-linux] [patch 15/23] Add cmpxchg_local to parisc
    ... non-SMP-safe counter that protects updates against interrupts. ... could be vastely used in the kernel. ... "Local atomic operations only guarantee variable modification atomicity ... that only one CPU writes to the local_t data. ...
    (Linux-Kernel)
  • Re: irqbalance mandatory on SMP kernels?
    ... compile your own kernel. ... utilization letting one CPU take them all. ... "Ooh, interrupts are hurting one ... Maybe the scheduler's load balancer ...
    (Linux-Kernel)
  • Re: Although CONFIG_IRQBALANCE is enabled IRQs dont seem to be balanced very well
    ... Probably kernel is optimizing this even with CONFIG option enabled. ... For a UDP bandwidth lower than 23 Mbps the interrupt rate at the receiver interface was lower than 2000 interrupts per second. ... In this case all interrupts were distributed to CPU 0. ...
    (Linux-Kernel)
  • Re: Interrupt using all the CPU
    ... Yes i suspected a bug when i first saw this because of what you say. ... 100% CPU usage should be noticed. ... In cases where interrupts reach 25%, the load is very easily noticable, ... is false, you wouldn't be able to log in to the box, the kernel itself would have trouble doing something, but interrupt handling. ...
    (freebsd-questions)
  • Re: disk speed regression kernel 2.6.29 and after
    ... When I installed kernel 2.6.30, ... I currently have 3 hard drives hooked up. ... ACPI: ... CPU: L2 Cache: 512K ...
    (Linux-Kernel)