[Spread-users] partition mask

[Weinsberg Yaron]

Hi,

There are 4 daemons running on the LAN.
Actually, I didn't get any Transitional messages so far -
I just wanted to know where it is configured.

Incrementing the timeout avoids transitional messages, but i had some
trouble solving the general
case (where there are partitions) :
1. Suppose there are 5 members in a group which need to ACK/NACK a message.
2. If a message is received after a transitional message, and a process is
not in a quorum it should leave the
group (which indicates an ACK response for the others).
BUT from what i have read (Idit Keidar,Yair Amir,Danny Dolev) a process
can't just look at the members list
(of a view change event) and decide if it is in a quorum because of possible
further partitions (and the danger in creating several primary components)

Does spread make it easier for a programmer or should i need to implement a
3PC in order to decide on a quorum ?
I really start to feel that using EVS is harder than VS (which supports
flush).... Am I right ?

    Thanks for your help!
        yaron.

btw, does the postgres-spread manager (which support database replication)
is open source?
where can i get it ?






----- Original Message -----
From: "Jonathan Stanton" <jonathan at cnds.jhu.edu>
To: "Weinsberg Yaron" <wyaron at sangate.com>
Cc: <spread-users at lists.spread.org>
Sent: Wednesday, March 13, 2002 6:31 PM
Subject: Re: [Spread-users] partition mask


> On Wed, Mar 13, 2002 at 05:28:14PM +0200, Weinsberg Yaron wrote:
> > Hi,
> >
> > I need some configuration tip:
> >
> > 1. I use several spread daemons in a special LAN configuration : Each
node
> > consists of two Ethernet interfaces
> > and a special switch is used in order to change the node interface when
a
> > failure occurs.
> >
> > 2. The switch exchange the interfaces in aprox. 2seconds.
> >
> > 2. How can I disable the detection of the temporary partition and avoid
a
> > transitional membership message?
>
> The token timeout that is configured by default is 5 seconds on single
site
> local networks and 10 seconds when multiple segments are configured. So
I'm
> not sure why a 2 second communication outage is triggering a membership.
>
> It could be that it isn't the time involved, but something else. Roughly
how
> many daemons are running on the lan? Also, when you do the switchover do
> both interfaces maintain the same ip address, or do they have separate ip
> addresses. Finally, when a switchover occurs do all of the daemons switch
to
> the new network, or only the one who had a failure.
>
>
> Jonathan
> --
> -------------------------------------------------------
> Jonathan R. Stanton         jonathan at cs.jhu.edu
> Dept. of Computer Science
> Johns Hopkins University
> -------------------------------------------------------
>