[Spread-users] Newbie question: why a limit of 128 daemons in a cluster?

[Jonathan Stanton]

I am resending this message because the server crashed while it was being
sent and I don't think anyone got it. If you did, please just ignore the
duplicate.

Two answers:

1) because of the model of reliable multicast with membership that spread
supports, there has to be some limit.
  
2) The reason it is 128 is because of a detail in the implmentation. When
Spread recovers from a network merge, one of the state messages only has
room for 128 machine's worth of state information.

3) In practice, depending on how you use it, the effective limit on the
number of active daemons is lower, the max I have heard of people using is
50 or so.
    
Yes, I can count and 3 != 2, but I came up with another reason while
writing.

Jonathan



On Tue, Sep 25, 2001 at 11:37:58AM -0500, Guerry Semones wrote:
> I'm having my first look at Spread and was curious why this restriction?
> 
> Is it a spread performance related restriction, or something in the
> router or network world that I'm not aware of?
> 
> Thanks!
> 
> 
> 
> 
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users

-- 
-------------------------------------------------------
Jonathan R. Stanton         jonathan at cs.jhu.edu
Dept. of Computer Science   
Johns Hopkins University    
-------------------------------------------------------