On Fri, Aug 16, 2002 at 11:36:09AM +0200, Leif Nixon ([EMAIL PROTECTED]) wrote:
> *However*, it turns out glibc already does this!
>
> sysconf(_SC_NPROCESSORS_CONF) should work on x86, sparc, alpha, mips
> and, probably, ia64.
>
Good tip, thanks! Much easier than doing it the hard way.
glibc e
leif-
that's a great tip! it looks like it works on cygwin and we had problems
with calling the number of CPUs on cygwin (well i guess i should say
matt rice was having problems since i'm still scratching my head amazed he
ported ganglia to cygwin in the first place).
thanks!
-matt
Today,
Leif Nixon wrote:
Preston Smith <[EMAIL PROTECTED]> writes:
On Thu, Aug 15, 2002 at 03:11:34PM +0200, Leif Nixon ([EMAIL PROTECTED]) wrote:
A tentative fix would be to instead count the number of lines
starting with "processor" in /proc/cpuinfo.
There's been a bug filed about
Preston Smith <[EMAIL PROTECTED]> writes:
> On Thu, Aug 15, 2002 at 03:11:34PM +0200, Leif Nixon ([EMAIL PROTECTED])
> wrote:
> > A tentative fix would be to instead count the number of lines
> > starting with "processor" in /proc/cpuinfo.
>
> There's been a bug filed about this exact issue on
On Thu, Aug 15, 2002 at 03:11:34PM +0200, Leif Nixon ([EMAIL PROTECTED]) wrote:
> A tentative fix would be to instead count the number of lines
> starting with "processor" in /proc/cpuinfo.
>
Done.
My cvs log:
Added mtu code for freebsd and aix, since it won't compile
without it there.
Pat
On Thu, Aug 15, 2002 at 03:11:34PM +0200, Leif Nixon ([EMAIL PROTECTED]) wrote:
> Hi,
>
> I was a bit surprised when Ganglia reported that a couple of my clusters
> have a grand total of zero CPUs.
>
> Seems monitor-core/gmond/machines/linux.c uses a buggy method
> to determine the number of proc
Hi,
I was a bit surprised when Ganglia reported that a couple of my clusters
have a grand total of zero CPUs.
Seems monitor-core/gmond/machines/linux.c uses a buggy method
to determine the number of processors in a x86 machine; it counts
the number of lines starting with "cpu" in /proc/stat and s