[rabbitmq-discuss] High-availability: Unsynchronised mirrors questions (3.0.2)

[Simon MacMullen]

See http://www.rabbitmq.com/ha.html#unsynchronised-slaves

And note that we will be making improvements in this area in 3.1.

Cheers, Simon

On 13/03/13 10:53, Alexis Van Der Merwe wrote:
> Hi there,
>
> I'm testing out the HA features using server version 3.0.2 and it seems
> to fulfill my requirements.
> There's one thing I don't really understand, and I was hoping someone
> would be able to enlighten me.
>
> I have a two (2) node HA cluster, for convenience, let's call them
> /NODE-MASTER/ and /NODE-SLAVE/.
>
>  1. I publish 100k messages to NODE-MASTER, and they are replicated to
>     /NODE-SLAVE/.
>  2. I take down /NODE-MASTER/ using *rabbitmqctl*.
>  3. NODE-SLAVE is promoted to master and retains the messages it
>     previously received.
>  4. Bring /NODE-MASTER/ back up and into the cluster
>
> This all works as I expect it to, but after the former master node
> rejoins the cluster,
> I see a message in the web-management console saying that one of the
> nodes is unsynchronised (see screenshot).
>
> Why does this occur, and more importantly, what does it mean in this
> scenario?
>
> Regards,
>
> A
>
> <https://lh6.googleusercontent.com/-bFcdaf1LZHU/UUBaaQiwPqI/AAAAAAAAA-0/JxUjiQWpbDc/s1600/unsynchronised_mirrors.jpg>
>
>
>
>
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>


-- 
Simon MacMullen
RabbitMQ, VMware