Re: Number of CPUs detected by Bind 9.4.2 on 4 CPU system running RedHat es 4.
On Tue, Apr 1, 2008 at 6:23 AM, Adam Tkac <email@example.com> wrote:[color=blue]
> On Mon, Mar 31, 2008 at 11:59:10AM -0600, greg kuechle wrote:[color=green]
> > Hello,
> > I have install bind 9.4.2 on a system with 4 CPUs running RedHat es4. I
> > compiled named with the --enable-threads and used the -n 4 flag when I
> > start named.
> > Mar 31 10:12:24 ******** named: starting BIND 9.4.2 -t /opt/named
> > -u named -n 4 -c /etc/named.conf
> > Mar 31 16:12:24 ******** named: found 1 CPU, using 4 worker[/color]
> > Before I upgraded the system it was running an older version of named[/color]
> > comes packaged with the OS.
> > Here is the output from the logfile.
> > Mar 24 11:34:13 ******** named: starting BIND 9.2.4
> > Mar 24 11:34:13 ******** named: using 4 CPUs
> > Did I compile named wrong?
> > Will BIND 9.4.2 use all of the CPUs if I use the -n 4 option at startup?
> > It looks like named is only running on one CPU. Will named start using[/color]
> > other CPUs once one CPU is up to 100% ?
> > Thank you for any help.
> I think you forgot mount /proc filesystem into chroot. You can try it
> with /proc mounted in chroot (for example $mount --bind /proc
> Adam Tkac, Red Hat, Inc.[/color]
Thank you Adam,
That did the trick. I mounted /proc in chroot and restarted named.
I am still using the -n 4 switch. The log output is:
Apr 1 14:01:58 dnsserver-1 named: found 4 CPUs, using 4 worker
When I run a ps -ef | grep name I only see one named process running. Is
this correct ?
I thought I would see 4 running.