I have used Ganglia for the last year or so in different cluster systems
(virtualized and non-virtualized). Never had a problem setting it up or
using it. For the last 5 days I have been trying to setup Ganglia in a
cluster made of 50 VMs runing RHEL5.3 without success...
Since i am not sure if the network support multicasting I am using the
unicast option. I have followed the instructions found in
http://wiki.freaks-unidos.net/ganglia-quickstart#using-unicast
To start I am trying to setup Ganglia on 5 VMs. I am following the same
setup in
http://www.mail-archive.com/[email protected]/msg02655.html
That is,
I have a masterHost running gmond and gmetad
gmond in masterHost is configured with udp-recv_channel of 8649 and
tcp-accept_channel of 8650. Bind to the hostname masterHost.
gmetad has been configured with one single data_source, more specifically:
data_source "my cluster" masterHost:8650
As per the instructions I configured gmond.conf in all the other nodes to
send data to masterHost. That is,
udp_send_channel {
host=masterHost
port = 8649
}
Everything seems to work fine till I bring up the fourth gmond (including
the one running on the masterHost). Then, gmetad web front reports as if
hosts are going down and up ... when in reality all hosts are up.
Can someone give me advice on how to debug this?
Thanks,
Suemar
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
