Re: N-way multi-master replication

[Adrien Futschik <adrien.futschik@atosorigin.com>]

Le mercredi 18 février 2009 16:27:43, Gavin Henry a écrit :
> ----- "Adrien Futschik" <adrien.futschik@atosorigin.com> wrote:
> > I am testing Multi-master replication with two masters.
> >
> > If I stop one of them and don't restart it before the end of retry
> > interval,
> > the replication process doesn't work anymore. (looks like it is
> > stopped).
> >
> > exemple :
> > - M1 & M2 are synchronised.
> > - stop M2
> > - add entries to M1
> > - wait for retry to fail (retry="5 5 300 5")
> > - restart M2
> > - previously added entries to M1 are replicated to M2
> > - modifying M2 (attributes) are not replicated to M1
> >
> > It look like M1 -> M2 : OK and M2 -> M1 : broken
> >
> > Is there a way to restart replication after configured retry are over
> > ? The
> > solution I found was to restart M1. After that everything seems to
> > work fine.
> >
> > I could set up retry so that the number of retry is very hight, but I
> > was
> > wondering if there was a way to restart synchronisation online.
>
> "If an error occurs during replication, the consumer will attempt to
> reconnect according to the retry parameter which  is  a list of the <retry
> interval> and <# of retries> pairs.  For example, retry="60 10 300 3" lets
> the consumer retry every 60 seconds for the first 10 times and then retry
> every 300 seconds for the next  3  times before stop retrying. The '+' in
> <# of retries> means indefinite number of retries until success.  If no
> retry was specified, by default syncrepl retries every hour forever."
>
> Why not just change it to retry="60 +" ?
 
I thought of that, I'll probably do something like that, but I was wondering 
if there was any other way to restart replication after retry ends.
A command like "start slave" for mysql. ...