Open Stack

My understanding is, in Rabbit versions prior to 3.x, the HA policy was
set client side when the queues were created. Hence the setting in the
config files. Since version 3.x, the paradigm changed and the ha policy is
now set server side, which is what you're seeing there.

Generally this is done by your automation when you setup rabbit and create
the policies.

On 9/1/15, 1:07 PM, "Ajay Kalambur (akalambu)" <akalambu at cisco.com> wrote:

>Hi
>How is the rabbit_ha_queues parameter used in configuration files like
>nova.conf, neutron.conf, cinder.conf etc
>
>
>What happens if on the rabbit node the ha queue is set to mirrored but
>the ha queues is set to False on client side
>[root at j10-controller-1 /]# rabbitmqctl list_policies
>Listing policies ...
>/ha-all all{"ha-mode":"all","ha-sync-mode":"automatic"}0
>...done.
>[root at j10-controller-1 /]
>
>
>
>
>
>I have rabbit_ha_queues=False
> set on nova and neutron.conf and from what I can see the queues seem to
>be mirrored. So why is this needed in nova.conf, neutron.conf etc
>
>
>This is Juno release
>
>
>