On Tue, 30 Dec 2003, Chris Johnson wrote:
> On Mon, 29 Dec 2003, Chris Johnson wrote:
>
> > Ok, hi all.
> >
> > Just got ganglia up. Using it to monitor a cluster. Using
> > ganglia 2.5.5.1 on RH7.3. Have the Web interface up. Have gmond
> > running on the master with gmetad and one compute engine. The master
> > has 2 CPUs the engine has 1. So far so good.
> >
> > Gstat and the Web interface both agree that the only thing they
> > see is the master node. Doesn't see the one compute engine.
> >
> > Obviously I'm doing something wrong.
> >
> > Can anyone give me and idea how his mcast_channel thing works and
> > what it should be set too? And yeah I read the rather spares
> > documentation. Need more.
> >
> > Any other ideas would be helpful. Thanks much. Happy new year.
> >
> >
>
> Hm, so I think I figured out that it's not how mcast_channel is
> set up. Netstat -a claims that is ESTABLISHED. Which means that
> gmetad and the gmond's SHOULD be talking to each other.
>
> Now a trafshow -N port 8649 shows that the master is talking to
> itself and the compute egines are talking to each other. The engines
> aren't talking to the master and visa versa. That would explain why
> the master isn't seeing the nodes; no traffic. Question is, why?
> Still no clue.
>
>
Got it! Amazing what sleep can do.
The master gmod wasn't using the interface the engine nodes were
on. Hence, no talking. Working now.
-------------------------------------------------------------------------------
Chris Johnson |Internet: [EMAIL PROTECTED]
Systems Administrator |Web: http://www.nmr.mgh.harvard.edu/~johnson
NMR Center |Voice: 617.726.0949
Mass. General Hospital |FAX: 617.726.7422
149 (2301) 13th Street |If ignorance is bliss, why aren't there more
Charlestown, MA., 02129 USA |happy people? Tea bag tag
-------------------------------------------------------------------------------
