Fedora Users — Re: irqbalance runs only once; is that right?

Hi Aaron,

On Tue, Aug 19, 2008 at 17:31, Aaron Konstam <akonstam@xxxxxxxxxxxxx> wrote:

On Tue, 2008-08-19 at 16:35 -0300, Andre Costa wrote:
> Hi,
>
> I just realized that, even though /etc/sysconfig/irqbalance has
>
> > >
> irqbalance runs only once and then bails out (at least it seems to be
> gracefully). This is what happens if I try to run it by hand:
>
> root:~# irqbalance --debug
> Package 0: cpu mask is 00000003 (workload 0)
> Cache domain 0: cpu mask is 00000003 (workload 0)
> CPU number 0 (workload 0)
> CPU number 1 (workload 0)
> Interrupt 20 (class ethernet) has workload 1
> Interrupt 8 (class timer) has workload 0
> Interrupt 0 (class timer) has workload 0
> Interrupt 17 (class storage) has workload 21
> Interrupt 21 (class storage) has workload 0
> Interrupt 16 (class legacy) has workload 123
> Interrupt 22 (class legacy) has workload 81
> Interrupt 1 (class legacy) has workload 1
> Interrupt 23 (class legacy) has workload 0
> Interrupt 19 (class legacy) has workload 0
> Interrupt 18 (class legacy) has workload 0
> Interrupt 9 (class legacy) has workload 0
> Interrupt 6 (class legacy) has workload 0
> Interrupt 4 (class other) has workload 0
>
>
>
> -----------------------------------------------------------------------------
> IRQ delta is 1892
> Package 0: cpu mask is 00000003 (workload 1154)
> Cache domain 0: cpu mask is 00000003 (workload 1153)
> CPU number 0 (workload 508)
> Interrupt 17 (storage/91)
> Interrupt 22 (legacy/412)
> Interrupt 23 (legacy/0)
> Interrupt 18 (legacy/0)
> Interrupt 6 (legacy/0)
> CPU number 1 (workload 645)
> Interrupt 20 (ethernet/10)
> Interrupt 21 (storage/8)
> Interrupt 16 (legacy/621)
> Interrupt 19 (legacy/0)
> Interrupt 9 (legacy/0)
> Interrupt 1 (legacy/0)
> Interrupt 4 (other/0)
>
> After this, it just quits with status 0, and /var/run/irqbalance.pid
> is removed.
>
> Is this normal behavior? (manpage says it should keep balancing IRQs
> each 10s if ONESHOT is not set)
>
> Version is irqbalance-0.55-9.fc9.x86_64
>
> Regards,
>
> Andre

That is interesting because the big leap forward in 2.6.x kernels is to
have scheduling lists that balance the work more efficiently among cpus.

You mean that irqbalance isn't really needed with 2.6 kernels? I'm not sure I need to run it, but since I am, it feels strange that it stops after the first run, even if (apparently) not told to do so...

-- 
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list