Thanks for the response. If possible i would like to reduce the amount of tooling. We
already have quite a lot. Do you think a single ovirt engine could cope?
Verstuurd vanaf mijn iPhone
Op 26 apr. 2016 om 14:19 heeft Alexander Wels
<awels(a)redhat.com> het volgende geschreven:
> On Tuesday, April 26, 2016 12:13:11 PM Joost(a)familiealbers.nl wrote:
> Hi all, i have a somewhat awkward requirement to deploy datacenters to
> around 500 satelite locations. Each datacenter will have 1 or 2 hypervisors
> and if 2 a cluster with glusterfs for shared storage will be deployed. I am
> expecting to run between 4 and 10 vms per data center. During operations
> it is expected 20% to 40% of the satelite locations will be down or have a
> very bad connection to the engine. I will still like to manage these
> datacenters from ovirt engine. Does anyone have any figures on how many
> datacenters i can add given what requirements for a single engine and
> perhaps some best practices for dealing with 'bad' connections between the
> hypervisors and the engine. Thanks a lot!
Sounds like a good use case for manage IQ. If possible I would do hosted
engines in each satellite location, and then use manage IQ to manage the
engines.
Alexander
ps.
Vrolijk Koning dag morgen
>
> _______________________________________________
> Users mailing list
> Users(a)ovirt.org
>
http://lists.ovirt.org/mailman/listinfo/users