Hi,
I have been using the Gmond Python metric modules (
http://sourceforge.net/apps/trac/ganglia/wiki/ganglia_gmond_python_modules)
to monitor network activity from virtual machines using the spoof host
functionality. I have followed the examples of spfexample.py and
spfexample.pyconf in the ganglia codebase. I'm using Ganglia 3.1.2 in
Ubuntu 10.04.
I created a list of spoof hosts by retrieving host names from libvirt 0.8.3
(using the python libvirt library -this list of spoof hosts gets dynamically
filled in the metric_init() function in spfexample.py). I have the problem
that whenever I start new virtual instances, they do not appear in ganglia
(no rrd files for the spoofed VMs) until I restart gmond on the host.
I know that normally, ganglia would not detect these new instances (as I'm
pulling data from libivirt - these new instances do not have gmond installed
on them). Has anyone else tried monitoring VMs like this or know of a
possible work-around (without having to install gmond on the VMs or restart
gmond on the host every time I start up a new machine?).
The same goes for when I kill an instance. My spoofed heartbeat function
shows that the instance is dead (I pull the state of the VM from libvirt
every 10 seconds), but the instance still shows as running in the ganglia
web portal after it's terminated (unless I restart gmond). Has anyone tried
this as well? Thanks!
-Bryan
------------------------------------------------------------------------------
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G. Thomson, author of the best-selling book "Blueprint to a
Billion" shares his insights and actions to help propel your
business during the next growth cycle. Listen Now!
http://p.sf.net/sfu/SAP-dev2dev
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
