[Spread-users] high cpu load

[George Schlossnagle]

Yair,

I believe if he has his ring set up as he does, it will always fail to
successfully muticast it's packet to it's neighbor and then have to unicast
it.  This would completely explain his retransmisions being the same as his
multicasts.

George

----- Original Message -----
From: "Yair Amir" <yairamir at cnds.jhu.edu>
To: <spread-users at lists.spread.org>
Cc: "Dirk Vleugels" <dvl at 2scale.net>
Sent: Friday, August 31, 2001 11:50 AM
Subject: Re: [Spread-users] high cpu load


> Hi,
>
> Instead of all that - lets have a monitor setting flow control as follows:
>
> Flow Control Parameters:
> ------------------------
>
> Window size:  60
>
>         cluster1        3
>         cluster2        3
>         cluster3        3
>         cluster4        3
>         loghost 3
>
> and see if there is any changes in terms of loses.
> You set the flow control and then you send it.
>
> WHat this does - it limits each daemon to 3 packets at most on every
round.
>
> Cheers,
>
> :) Yair.
>
> Michael Stiller wrote:
> >
> > Hi George,
> >
> > i'm working with Dirk on the same problem.
> >
> > Does it matter if we configured spread to use a multicast group which is
> > on another interface as the ip addresses we used in the config file ?
> >
> > E.g.:
> >
> > Spread_Segment 224.0.224.1:4803 { # Multicast route is on eth1
> >         cluster1        192.168.100.1 # IP Adress of eth0
> >
> > In the beginning we tried to use the "correct" ip address of
> > eth1 but spread failed to run. Unfortunately i don't remember
> > the error messages anymore.
> >
> > Cheers,
> >
> > -Michael
> >
> > _______________________________________________
> > Spread-users mailing list
> > Spread-users at lists.spread.org
> > http://lists.spread.org/mailman/listinfo/spread-users
>
>
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users
>