I am using gmond 3.1.2 on a Solaris 10 SPARC cluster
After some time, it seems that gmond just becomes deaf.
I can connect to the gmond TCP port (8649) and get the XML, but all the metrics
are expired (TN > TMAX). All the values for TN are aproximately the same for
all metrics, including metrics from other hosts in the same cluster. My
conclusion is that gmond is not listening anymore for metrics on the multicast
channel, it just ignores any broadcasted metrics. In a single word, it is deaf
on the multicast channel.
I can connect to another gmond node for some time and get the metrics, but
eventually, it also has the same symptoms. (TN > TMAX for all metrics,
including other nodes of the same cluster). Eventually all my nodes are no
longer receiving any metrics. All of them report the metrics as expired (TN >
TMAX)
Does anyone has experienced the same problem?
Any suggestions, solutions?
In addition to the standard metrics, my application broadcast its own specific
metrics (like #requests/minute, % JVM free memory, amont others) using the same
XDR protocol instead of calling gmetric. (In other words, I "embedded" gmetric
into my application). Would this affect gmond ? (Of course, my app does not
broadcast the standard metrics)
After gmond becomes deaf (on the multicast channel), Ganglia Web reports my
cluster as being "down".
-Jorge
------------------------------------------------------------------------------
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
