Hi,
At first everything worked fine, but after a while I get the following
message for some clusters:
Cannot find any metrics for selected cluster "cube28", exiting.
Check ganglia XML tree (telnet 127.0.0.1 8652)
This is from the webfrontend. If I check the ganglia XML tree as suggested,
I do find a cluster "cube28", with a couple of node tags.
The grid monitors 2 clusters "cube28" & "cube83". Looking at the ganglia
XML tree, I can see that the cube83's "reported" timestamp is lagging some
12000 (seconds/milliseconds ?) as compared to cube28's "reported" timestamp.
Can this have anything to do with the problem ? All nodes within a cluster
have their clocks synch'ed to one another, but these clocks aren't synch'ed
between clusters. Do all clusters (& the host running gmetad) need to be
time-synch'ed ?
Cheers,
Arnaud
--
Arnaud Installe
<[EMAIL PROTECTED]>
Senior Software Engineer EMC Belgian Development
Group
