Hello,

I went through all the archives of this list and also posted to
ganglia-general and and got some hints from there but I am still having
some problems getting ganglia to work correctly.
I am hopping someone here can maybe point me in the right direction.

For now I don't really have any "clusters" My network consists of
several Single SMP servers which I would like to have show up on ganglia
as a "grid".

All my servers are on the same network or subnet.

host "venus" SGI 8CPU server running gmod only not gmetd
here is its gmond.conf

venus 35# cat /etc/gmond.conf
name "venus SGI Server"
owner "ccs.tulane.edu"
trusted_hosts 129.81.120.69
setuid     nobody
mcast_channel 239.2.11.79
mcast_port    8655
mcast_if  ef0
xml_port     8655
num_nodes 1
debug_level 0




host "mac" 2CPU apple G5 running gmond only not gmetad
here is its gmond.conf
[mac:~]% cat /etc/gmond.conf
name "mac Apple G5"
owner "ccs.tulane.edu"
trusted_hosts 129.81.120.69
num_nodes 1
mcast_channel 239.2.11.83
mcast_port    8656
xml_port     8656
setuid     nobody
debug_level 0



Sgi server running gmetad
here is its gmetad.conf file

ccs1 50# cat /etc/gmetad.conf
data_source "venus" venus.ccs.tulane.edu:8655
data_source "mac" mac.ccs.tulane.edu:8656
gridname "Tulane"
scalable on
debug_level 0
authority "http://ccs1.ccs.tulane.edu/Grid/stats";
rrd_rootdir "/var/ganglia/rrds"
trusted_hosts 127.0.0.1
setuid_username "nobody"



If you look at the ganglia page here
http://www.ccs.tulane.edu/Grid/stats/

there are a couple of things wrong with it
1) the "CPUs total:" "Hosts up:" and "Hosts down:" are not being set
2)if you choose a source or click on the graph for one of the servers
say the SGI box you get taken to a page where the
"CPU Total: 10" this is not true the SGI box only has 8 CPUS

I think it is adding the "apple mac" 2 CPUS to this total.

Also the "mac" shows up as a node of the SGI machine in the pull down
menu that states "--Choose a Node".


For some reason ganglia is "Clustering" my two servers and not keeping
them as two separate machines.

Any ideas as to what might be the problem?

Thank you
Rene


Reply via email to