[erlang-questions] massive distribution
Tue Dec 1 15:22:06 CET 2009
On Tue, 2009-12-01 at 09:08 -0500, Kevin A. Smith wrote:
> Fully connected meshes suck for large numbers of nodes. Erlang provides a number of
> knobs to control how a cluster is stitched together such as "-connect_all false"
> and "-hidden".
Which would entail keeping track of connected nodes and connection
> Also, tuning the net tick time (see man 3 net_kernel and man 6 kernel) can be helpful
> in keeping a large cluster running.
I fiddled around with those a bit. I don't have the exact values at
hand, but I set net_ticktime to rather large values, something like
300s, without substantial improvements in the number of nodes able to
keep a stable connection.
> On Dec 1, 2009, at 5:32 AM, Roberto Ostinelli wrote:
> >> If you're talking about the built-in Erlang distribution mechanisms --
> >> FWIW: in a quick test I did a while ago I had trouble keeping stable
> >> connections for more than ~80 connected nodes. I must admit I was a bit
> >> surprised as the "Efficiency Guide" seems to imply that the number
> >> should be much higher (?).
> >> Possibly the number of known (but not connected) nodes can be higher
> >> than this.
> >> HTH,
> >> peter.
> > thank you peter.
> > 80 connected nodes seems a low number if one seriously need to build
> > cloud applications.
> > i guess that one needs to develop his own mechanisms for
> > interconnecting nodes in a way more similar to a custom 'mesh
> > networking' [pardon the conceptually wrong extension of this term]?
> > would that be more appropriate, to your belief?
> > cheers,
> > r.
> > ________________________________________________________________
> > erlang-questions mailing list. See http://www.erlang.org/faq.html
> > erlang-questions (at) erlang.org
More information about the erlang-questions