On Fri, 12 Apr 2002, Bill Denney wrote:
> On Fri, 12 Apr 2002, Bert Freudenberg wrote:
> > I saw this on my servers, too. We never figured out exactly what was
> > wrong, but I think it is due to some kernel/glibc issue. The exact
> > same squeak executable runs fine now. Are you running current
> > versions? And what VM are you using?
>
> This morning I upgraded the system from kernel 2.4.9 to 2.4.18 and that
> dropped the processor usage down to ~40% (both kernels were redhat
> compiled-- not custom because I need to leave this box having very easy
> administration). I then upgraded gcc/glibc and gcc3/glibc3 to the most
> current rawhide versions and that brought the processor usage up to ~75%.
>
> I guess that this is better, at least the processor isn't pegged at 100%
> anymore. Thanks for all the help.
>
> I don't know if there is a "known issues" page anywhere, but if there is
> then it would probably be a good idea to recommend upgrading away from the
> stock redhat 2.4.9 kernel (and perhaps any 2.4.9 kernel) somewhere.
We can't be sure about this until we really found the problem. You could
try profiling, for example, to see where exactly the time is spent. I'm
running red-hat stock kernel/glibc combo, and usage is 0%:
kernel-smp-2.4.9-31
glibc-2.2.4-19.3
-- Bert
