Paul Belanger <[email protected]> writes:

> So, I was thinking about this the other day, what would be needed to make
> zuul-server more redundant.  Since we can scale out merges, why not the server
> part too.  This should help mitigate issues (eg: memory leaks, crash).
>
> I was thinking something like memcache / redis / some other tool to 
> externalize
> queues would be needed.  But obviously, this adds additional overhead to 
> running
> zuul.
>
> It would be interesting to see if ansible-launchers could be scaled out too 
> over
> multiple hosts. Since we'd be pushing more SSH connections from zuul now.

The launchers are absolutely a scalable component.

-Jim

_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra

Reply via email to