Open Stack

Hi,

I'm not sure if you used any deployment tool for your environment, but
usually for such failover there's a load balancer in conjunction with VRRP
or Anycast, that able to detect when controller1 is down and forward
traffic to another controller.

As example, OpenStack-Ansible as default option installs haproxy to each
controller and keepalived to implement VRRP and Virtual IPs. So when glance
is down on controller1, haproxy will detect that and forward traffic to
other controllers. If controller1 is down as a whole, keepalived will
detect that and raise VIP on controller2, so all client traffic will go
there. Since controller2 also have haproxy, it will pass traffic to
available backends based on the source IP.


пт, 9 дек. 2022 г., 09:06 pradeep <pradeep8985 at gmail.com>:

> Hello All,
>
> I understand that glance-api container run only one controller always. I
> have tried to check if the glance-api container to switch over to the next
> available controller by means of rebooting and stopping the containers, but
> nothing happened. Is there a way to make sure that glance-api container
> switches to other controllers if controller 1 is not available.
>
> Thanks for you help
>
> Regards
> Pradeep
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.openstack.org/pipermail/openstack-discuss/attachments/20221209/7ba9616d/attachment.htm>