Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-31 Thread Martin Knoblauch
--- Joel Krauska [EMAIL PROTECTED] wrote: Rick Mohr wrote: The unicast approach does save on gmond memory usage as you mentioned. It's up to each site to determine just how much memory the metrics will take up, and if it is considered a significant amount. (But it can get

Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-31 Thread Joel Krauska
Martin Knoblauch wrote: just in case you did not know: http://ganglia.sourceforge.net/gmetric/ Hadn't known about this -- thanks. Question: I just went an covnerted to using UDP unicasts. The gmetric man page seems to imply that it only supports the multicast comm method. Is there a way

Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-31 Thread Ramon Bastiaans
The description may be a bit vague. The thing that gmetric does is insert the values into the running local gmond's xml. Then it's up to the gmond('s config) how it is sent, be it unicast or multicast. Joel Krauska wrote: Martin Knoblauch wrote: just in case you did not know:

Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-31 Thread Martin Knoblauch
Joel, 2gmetric (at least in 3.0.x) takes a -c argument where you can specify the path to gmond.conf. gmetric will then use any transport defined for gmond. Simple, isn't it? Martin --- Joel Krauska [EMAIL PROTECTED] wrote: Martin Knoblauch wrote: just in case you did not know:

Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-31 Thread Ben Hartshorne
On Tue, Jan 31, 2006 at 12:15:19AM -0800, Martin Knoblauch wrote: just in case you did not know: http://ganglia.sourceforge.net/gmetric/ Everyone is invited to contribute to the repository. Martin, I believe someone else has pointed out that submissions have been closed (and

Re: [Ganglia-general] Pointers on architecting a largescale ganglia setup??

2006-01-30 Thread Joel Krauska
Rick Mohr wrote: The unicast approach does save on gmond memory usage as you mentioned. It's up to each site to determine just how much memory the metrics will take up, and if it is considered a significant amount. (But it can get somewhat big on a large cluster like mine with a bunch of