Bernard and all,
Thanks for your helpful suggestion! I've tried to run gmond on different ports
on a single server and it does work.
But could there be any chances of step further -- I mean, each client sends
out UDP packets with same, or different cluster name included; the single
server runs a single gmond listening on a single port to assemble the packets
and hand them to gmetad without cluster name strpping off or rewrite as stated
in the gmond.conf:
...
/* If a cluster attribute is specified, then all gmond hosts are wrapped inside
* of a <CLUSTER> tag. If you do not specify a cluster tag, then all <HOSTS>
will
* NOT be wrapped inside of a <CLUSTER> tag. */
cluster {
name = "greenstreet-cluster"
...
If gmond will send/receive, keep and handover the different cluster names in
multicast/unicast UDP packets, then it will make ganglia a perfect tool in
virtual computing cloud -- where the proxy head nodes is impossible to get,
every node is equally dynamic and virtual.
Thanks.
--- On Tue, 9/30/08, Bernard Li <[EMAIL PROTECTED]> wrote:
> From: Bernard Li <[EMAIL PROTECTED]>
> Subject: Re: [Ganglia-general] possible ganglia setup without proxies in
> computing cloud?
> To: [EMAIL PROTECTED]
> Cc: [email protected]
> Date: Tuesday, September 30, 2008, 3:47 AM
> HI David:
>
> On Mon, Sep 29, 2008 at 5:18 PM, Robinson Tiemuqinke
> <[EMAIL PROTECTED]> wrote:
>
> > I am a newbie to ganglia setup. I am pretty
> interested in ganglia for its power and hierarchy like
> setup: Grid -> Cluster(s) -> node(s).
> >
> > I've read the wiki at
> http://www.ibm.com/developerworks/wikis/display/WikiPtype/ganglia
> which is pretty helpful. But the problem of the setup is
> that proxies nodes with gmond/gmetad are in need. The
> central gmetad server talks with the proxies with
> 'datasource *' statement setup. But I like a setup
> without proxies -- all nodes from a handful of clusters(each
> node is setup with only one <cluster> section for
> sure) send UDP packets directly to the only gmond/gmetad
> server.
> >
> > The situation for me is: my environment is in
> completely virtual computing cloud, there are >20
> clusters/hostgroups but all nodes are dynamic/virtual and so
> proxy (head)nodes are impossible to fix -- they are all
> dynamic and can die hard any minute.
> >
> > Any one can shed a light on this? Thanks a lot.
>
> What some people do is run multiple gmonds on the central
> gmetad
> server each talking to different cluster gmonds. That way,
> your
> gmetad can query the gmond running on localhost (but via
> different
> ports) using the data_source directive.
>
> Would this work for you?
>
> In a dynamic environment, remember to set TTL for hosts so
> that when
> they die, they are automatically pruned from the list of
> hosts in a
> cluster.
>
> Cheers,
>
> Bernard
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
