Hi Manuel: Can you please clarify whether you are trying to setup Ganglia with unicast or multicast?
You can also get more troubleshooting information by running gmetad/gmond in debug mode (-d 2) and looking at your apache error logs. Cheers, Bernard On Tue, Dec 21, 2010 at 12:19 PM, Tarabas <[email protected]> wrote: > Hi Bernard, > > well it seems I was to quick to celebrate ... I just happened to get > the webinterface to show the clusters as a grid-overview only once and > only by chance as it seems. > > After reloading, the same behaviour as before occurs, that > it seems that the web-interface mixes up the clusters. > > I now configured the gmond's to 8649 (local server with gmond and > gmetad) and 8660 (gmond remote) as well as 2 more gmond's at 8661 and > 8662 > > You can observe the behavior at: > http://212.91.236.254/ganglia/ > > The following address is a sample of a cluster that gets metrics but > does not show in the list and seems to mix up the hostnames with the > first Cluster, that is the only one of 4 clusters that is working: > > http://212.91.236.254/ganglia/?c=eva&m=cpu_report&r=hour&s=descending&hc=4&mc=2 > > Do you have any other advice I can try to fix this web-frontend problem? > > Best regards > Manuel R. > > Tuesday, December 21, 2010, 8:33:29 PM, you wrote: > > BL> Hi Manuel: > > BL> This is a common misconception of Ganglia. In order for hosts to show > BL> up as different clusters, they will need to cluster on different > BL> ports. In your case, you can try changing gmond.conf for server-B to > BL> use port 8650 instead. Re-start gmond and then point gmetad to that > BL> port as well and you should have 2 clusters instead of one. Remember > BL> not to use ports 8651 and 8652 as these are gmetad reserved/default > BL> ports. > > BL> Good luck, > > BL> Bernard > > BL> On Tue, Dec 21, 2010 at 8:53 AM, <[email protected]> wrote: >>> Hello, >>> >>> I have a strange problem with my fresh ganglia install. >>> >>> I installed gmond/gmetad on one Server (server-A) >>> I installed gmond on another Server (server-B) >>> >>> Gmond-Config on "server-A" is: >>> >>> cluster { >>> name "server-A" >>> [...] >>> } >>> >>> udp_send_channel { >>> host=<ip-of-server-A> >>> port = 8649 >>> ttl = 1 >>> } >>> >>> udp_recv_channel { >>> port = 8649 >>> bind = <ip-of-server-A> >>> } >>> >>> tcp_accept_channel { >>> port = 8649 >>> bind = <ip-of-server-A> >>> } >>> >>> Gmond-Config on "server-B" is: >>> >>> cluster { >>> name = "server-B" >>> [...] >>> } >>> >>> udp_send_channel { >>> host=<ip-of-server-B> >>> port = 8649 >>> ttl = 1 >>> } >>> >>> udp_recv_channel { >>> port = 8649 >>> bind = <ip-of-server-B> >>> } >>> >>> tcp_accept_channel { >>> port = 8649 >>> bind = <ip-of-server-B> >>> } >>> >>> Gmetad-Config on "server-A" is: >>> >>> data_source "clusterA" <ip-of-server-A>:8649 >>> data_source "clusterB" <ip-of-server-B>:8649 >>> >>> Now the sending seems to work, server-A and server-B get queried by the >>> gmetad and the rrd files are created according to the expected cluster >>> configuration. I verified this by telnetting to the gmond's on the two >>> machines and to the gmetad. Both Clusters are shown correctly in the >>> gmetad-xml and all hosts of the clusters (always just one host) are also >>> shown correctly. >>> >>> When i now open the Web-Interface, I get the server-B-Cluster preselected >>> and cannot change the cluster. In the hosts dropdown I can see the one host >>> for the server-B cluster correctly. >>> >>> When I manually change the url I will get correct results if I use >>> "c=server-A" and "h=host.for.server-A", since the gmetad correctly created >>> the RRDs. But the bottom graph on the page is broken in this case and >>> shows "c=server-A&h=host.for.server-B" which of course will not work. >>> >>> It somehow seems, that the web frontend cannot correctly read the >>> information from the gmetad in some way and cannot render it correctly >>> although when i look at the xml, it seems correct. >>> >>> There are no Errors in either gmond or gmetad shown at debug level and i >>> will only get an error in the webserver logfile, that the rrd for the wrong >>> host cannot be found. >>> >>> Can anyone help me with this strange problem?! >>> >>> Best regards >>> Manuel R. >>> >>> ------------------------------------------------------------------------------ >>> Forrester recently released a report on the Return on Investment (ROI) of >>> Google Apps. They found a 300% ROI, 38%-56% cost savings, and break-even >>> within 7 months. Over 3 million businesses have gone Google with Google >>> Apps: >>> an online email calendar, and document program that's accessible from your >>> browser. Read the Forrester report: http://p.sf.net/sfu/googleapps-sfnew >>> _______________________________________________ >>> Ganglia-general mailing list >>> [email protected] >>> https://lists.sourceforge.net/lists/listinfo/ganglia-general >>> > > > > Gruss ... Manuel ... > > ------------------------------------------------------------------------------ Forrester recently released a report on the Return on Investment (ROI) of Google Apps. They found a 300% ROI, 38%-56% cost savings, and break-even within 7 months. Over 3 million businesses have gone Google with Google Apps: an online email calendar, and document program that's accessible from your browser. Read the Forrester report: http://p.sf.net/sfu/googleapps-sfnew _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
