Bernard Li wrote:
> Hi Daniel:
>
> On Tue, Nov 10, 2009 at 8:30 AM, Daniel Pocock <[email protected]> wrote:
>
>   
>> Which gmetad is intended to be on the future roadmap?
>>
>> For a large site, do you believe it is fair to say that the C
>> implementation is best for performance?
>>     
>
> I think if possible, we should perform a side-by-side performance test.
>
> It should be possible to run both gmetad and gmetad-python polling the
> same sets of gmonds for data.  The gmetads can reside on different but
> similar servers.
>   
I've been meaning to do something like this anyway, with a view to 
finding out how scalable Ganglia is when handling a large number of 
hosts and metrics at short polling intervals.

One related issue for me is distributing the load across multiple gmetad 
servers dynamically - is something like this on the roadmap for either 
of the gmetad versions?  At the moment, the administrator has to manage 
the load and move things around, but with users feeding their own 
metrics in, the need for Ganglia to adapt dynamically and distribute the 
load becomes more important.

My first thoughts on the matter involved some kind of shared storage 
containing all the RRDs, and using DLM to make sure that only one gmetad 
can write the RRDs for a particular host or cluster.



------------------------------------------------------------------------------
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-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-developers

Reply via email to