El mar, 14-05-2002 a las 17:53, david losada carballo escribió: > Hi Carlos, >
Thanks a lot David. The problem was in fact the multicast direction. I think I must take a look to my old networking knowledges... Just with the default multicast IP it works perfectly, and now I see all the computers inside the cluster, looks pretty Moltes gràcies i bona tarda !! > I think that the problem you have here is that you're not using the right > IP addresses. Ganglia monitors use a IP multicast address (default is > 239.2.11.71, can changed via the -mcast_channel, as you have noted). > 10.x.x.x addresses are not multicast, a multicast address is one in the > range 224.0.0.0-239.255.255.255. > > I suggest a couple of solutions: > > 1) add in each host a route for 239.2.11.71 (or whatever IP in the > multicast range you want to use) that goes to the management network > interface > route add -host 239.2.11.71 eth1 > > 2) add a route for the whole multicast range of addresses towards the > management interface: > route add -net 224.0.0.0 netmask 240.0.0.0 eth1 > > hope this helps > > kind regards, > > On 14 May 2002 17:33:23 +0200 > Carlos Javier Ramos <[EMAIL PROTECTED]> wrote: > > > Hello ! > > > > Im trying to monitorize a LVS cluster using ganglia. > > > > Each machine in the cluster are connected to different networks. > > > > Master Load balancer 1 > > - intranet network ( <- default route) > > - data network > > - management network > > > > Slave Load balancer 2 > > - internet (only active if Master is down!) > > - data network (only active if Master is down!) > > - management network > > > > > > Real Server 1 (Master Load balancer data network is default > > route)- data network > > - management network > > > > Real Server 2 (Master Load balancer data network is default > > route)- data network > > - management network > > > > > > Monitoring Box > > - management network > > - intranet network. ( <- default route) > > > > > > I have installed gmond, auth, gexec in all machines, and also necessary > > tools to do web monitoring through Monitoring Box, rrdtool, > > ganglia-php-rrdtool... > > > > By default i see that gmond has used the intranet network as the network > > to do monitoring, (i suppose due to this machines have the default route > > set to the intranet network)in the graphics (i see uptadated graphics, > > php works...) > > > > - Monitoring Box > > - Load Balancer 1 > > > > > > But i want gmond _to use the management network_ in order to see the > > other machines,so I edit an /etc/gmond.conf and use > > > > mcast_channel 10.10.20.1 > > > > where 10.10.20.1 is the IP of the Monitoring Box into management > > network. > > > > it did not works, so I add a default route to host in each computer > > > > route add -host 10.10.20.1 eth1 > > > > where eth1 was the interface connected to the management network, but > > instead of all it only show computers connected to the intranet. > > > > There is a simple hub between all the machines, no strange routers... > > > > Where i can specify gmond daemon to look for boxes in management network > > but not in our intranet? > > > > If I use gstat in the Monitoring Box I recevie the 2 intranet ip of the > > machines: > > > > CLUSTER INFORMATION > > Name: unspecified > > Hosts: 2 > > Gexec Hosts: 2 > > Dead Hosts: 0 > > Localtime: Tue May 14 16:38:49 2002 > > > > CLUSTER HOSTS > > Hostname LOAD CPU > > Gexec > > CPUs (Procs/Total) [ 1, 5, 15min] [ User, Nice, System, Idle] > > > > 192.168.1.207 > > 1 ( 0/ 162) [ 0.00, 0.00, 0.00] [ 0.5, 0.0, 0.1, 98.9] > > ON > > 192.168.1.208 > > 1 ( 0/ 133) [ 0.18, 0.24, 0.18] [ 37.6, 0.4, 0.3, 61.6] > > ON > > > > The default route in this machine is a fw located in 192.168.1.0 > > network. > > > > i also set in /etc/profile this lines: > > > > export GEXEC_GMOND_SVRS="gpphost1 gpphost2 director1 director2 doctor" > > export GEXEC_SVRS="gpphost1 gpphost2 director1 director2 doctor" > > export GEXEC_SVRS="gpphost1 gpphost2 director1 director2 doctor" > > > > where "gpphost1 gpphost2 director1 director2 doctor" is the name of the > > machines listed above. > > > > How can I specify the management network to do monitoring? > > It could be a routing problem? > > > > Thanks a lot in advance. > > > > -- > > -------------------------------------------------------- > > Carlos J. Ramos Genasys II Spain, S.A.U. > > Administrador de Sistemas Ventura de la Vega 5 > > Tel:+34-91-3649100 ext 170 28014 Madrid > > Fax:+34-91-3649119 Spain > > Email: [EMAIL PROTECTED] > > -------------------------------------------------------- > > > > > > _______________________________________________________________ > > > > Have big pipes? SourceForge.net is looking for download mirrors. We > > supply the hardware. You get the recognition. Email Us: > > [EMAIL PROTECTED] > > Ganglia-general mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/ganglia-general > > > -- > David Losada Carballo <[EMAIL PROTECTED]> > Tècnic de Suport a la Docència Informatitzada > Dept. d'Enginyeria Mecànica. Escola Tècnica Superior d'Enginyeria Química. > UNIVERSITAT ROVIRA I VIRGILI. TARRAGONA. > Tel. +34 977 559636 Fax +34 977 559691 > Avinguda dels Països Catalans, 26 > Campus Sescelades > 43007 TARRAGONA. SPAIN. -- -------------------------------------------------------- Carlos J. Ramos Genasys II Spain, S.A.U. Administrador de Sistemas Ventura de la Vega 5 Tel:+34-91-3649100 ext 170 28014 Madrid Fax:+34-91-3649119 Spain Email: [EMAIL PROTECTED] --------------------------------------------------------
