On Thu, Mar 13, 2008 at 03:09:01PM -0500, Martin Hicks wrote:
>
> (send to ganglia mailing list also)
>
> On Thu, Mar 13, 2008 at 12:19:55AM -0700, eliott wrote:
> > > I think I want to run gmetad on each head node, and to use that RRD data
> > > without
> > > regenerating it on the admin node. Is that possible?
> >
> > first you might try running gmetad on the head notes, and gmetad on
> > the admin node, and having the admin gmetad pull from the head gmetad
> > instances. Configure the head nodes for 'scalability mode'.
> >
> > like the 'yellow' and 'green' clusters in this image:
> > http://www-941.ibm.com/collaboration/wiki/download/attachments/3181/ganglia_cluster.gif
> >
> > image from:
> > http://www-941.ibm.com/collaboration/wiki/display/WikiPtype/ganglia
> >
> > I have never done the above myself (haven't needed to yet), but you
> > might try it before moving to custom rrd sharing mechanism.
>
> The information seems to indicate that you don't need the web server on
> the "green" cluster, but do you?
After more playing today, I've convinced myself that you do.
> I tried to set this up, although I had some difficulties.
>
> I have the admin node pulling from the head nodes on 8651 (gmetad), but
> I don't get my cluster names showing up. Debug output on the admin node
> from gmetad looks like:
>
> [Rack 1] is a 2.5 or later data stream
> Found a <GRID>, depth is now 1
> Found a </GRID>, depth is now 0
> Writing Summary data for source unspecified, metric disk_free
> Writing Summary data for source unspecified, metric bytes_out
> Writing Summary data for source unspecified, metric proc_total
> Writing Summary data for source unspecified, metric cpu_nice
^^^ This is the gridname of the
gmetad running on the head nodes.
>
> So it knows that the source is "Rack 1", but then below it is saying
> "source unspecified".
>
> In the web interface I now have a grid called "unspecified" and clicking
> those images seems to try to redirect me to a webserver running
> on the head nodes (which isn't installed...)
>
> This leads back to the initial question: Do I need the PHP web interface
> installed on all of the "Green" clusters (as per the IBM images, above)?
So, I think the IBM example above is not set up the way the author
expects. On his yellow cluster's gmetad he specifies:
data_source "green" green23
and claims that this will pull from gmetad on green23. I believe that
this pulls from gmond, and as such it also writes all the RRD data to
the yellow machine.
When I specify:
data_source "green" green23:8651
I can see the green Grid in the web interface, but it redirects all
requests for more information to http://green23/ganglia
This won't work for me because my green clusters aren't accessible
externally.
mh
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
