On Mon, 15 Jan 2007, matt massie wrote:
> On Thu, 2007-01-11 at 17:31 -0800, Matthias Blankenhaus wrote: > > Hello ! > > > > I have read through the Paper "The ganglia distributed monitoring > > system..." (Massie, Chun, Culler) from 2004 and found lots of > > extremely valuable information. > > > > I am concerned about the virtual memory utilization by gmond / > > gmetad based on data provided by the paper. However, the paper also > > hints at Ganglia design flaws, e.g. using more threads than necessary, > > and proposes alternatives like I/O multiplexing or reducing the > > thread stack size. > > > > My questions: > > > > (01) Is the data provided by this excellent paper still > > up to date or are the number different (better) ? > > the ganglia monitoring daemon (gmond) has been improved since that paper > was written. gmond now uses i/o multiplexing (no threads!) and lighter > data structures for storing data. you can easily run "top" to see just > how light it really is in your deployment... > Thanx, Matt. This is good news. Shortly I will run Ganglia on a test cluster with growing size and then be able to observe the values. I will make them available online. > > (02) Have the design issues been resolved in the meantime ? > > yes. however, gmetad still uses threads and it a bit of a beast but > that's ok. it's not meant to be light-weight or installed on every > machine (just the machine serving web pages). > Ditto. And yes, I am less concerned about the gmetad as it will run on a subset of machines. > -- > matt massie > phone: 415.692.0828 x2843 > fax: 415.278.0441 > http://archrock.com/ > >
