>>> On 10/18/2007 at 7:16 PM, in message <[EMAIL PROTECTED]>, Klaus Steden <[EMAIL PROTECTED]> wrote:
> Hello, > > I am investigating implementing some custom metrics in ganglia for use with > greceptor, specifically, to monitor hardware attributes. > > It looks straightforward, but I think I've hit a bit of a wrinkle. I've got > systems in my cluster that are slightly assymetrical (i.e. some have two > CPUs, some have only one CPU) -- which means that I would need to make two > copies of my very simple greceptor plugin -- one for CPU 0, and one for CPU > 1, and hardwire the specific CPU information into each one. > > This is not really ideal, as that means I have to do a hardware inventory on > a system-by-system basis to avoid creating more scripts than CPUs, or fewer > scripts than CPUs. > > What would be ideal would be if I reference CPU information as generic > variables in the plugin, and it would be able to call each CPU individually > and return numbers to ganglia, i.e. > > -- cut -- > def value(self): > cmd = 'read-cpu-temp %i' % cpu_id > temp = os.popen(cmd).readline() > return int(temp) > -- cut -- > > The question is ... is it possible to pass the CPU identity into the > > > > Is such a thing possible? > > I looked through Google and the ganglia-info list archives and couldn't find > any mention of it. Is there some documentation on greceptor plugins? > If you are talking about creating a python metric module then the answer to your question is 'yes'. In the metric_init() function, you will need to discover the hardware and then dynamically create a metric descriptor for each element. Then within your metric_handler() function, determine which element is being requested and return the value. Take a look at the multidisk.py module at http://ganglia.svn.sourceforge.net/viewvc/ganglia/trunk/monitor-core/gmond/python_modules/disk/multidisk.py?revision=846&view=markup . It discovers all of the hard drives on the box and dynamically creates a metric descriptor for each one. Then it returns various metrics depending on the drive metric that is requested. Brad ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
