Hi,I looked at , and it isn't clear about which host / hosts the EIPs will get mapped to. This makes a difference in the cost of my solution. If I have 10 EIPs and 10 hosts, and I map one per host, the ip addresses are free. But if OpenShift needs me to map them all to one host, I need to pay for 9 (because the first one is free). How does this work? To be clear, I'm asking about v3.
I'll ask this and other v4 questions in the forum you pointed me to.Thanks,Marvin_______________________________________________On Wed, Mar 6, 2019 at 3:53 PM W. Trevor King <wking redhat com> wrote:On Wed, Mar 6, 2019 at 12:38 PM Just Marvin wrote:
> Firstly - is this the right place to ask questions pertaining to the v4 developer preview? If not, would appreciate a pointer to the right place please.
 suggests  for the v4 developer preview.
> I have questions pertaining to how one would install on AWS (v4 or v3). AWS charges for Elastic IP's that are not mapped to a running EC2 instance, or additional elastic IP's (more than one) mapped to a EC2 instance. I know how many elastic IPs I need for routes, but I'm not sure how these IPs need to be assigned. Do I say that all those IP addresses are going to be assigned to the host running the SDN (openvswitch?) components? Or do I distribute them across the worker nodes? Do I need an elastic ip address for kube-dns (and if so, which host is that on - the master)?
The v4 installer  will handle all of this for you. We have
documentation for user-supplied infrastructure on AWS in the pipe, but
nothing I can link yet. Docs for the current EIP allocation are in
> If I need to enable https end-user traffic, will I need to install CA (but private CA) cert generating components, or does Openshift have the capability to dynamically generate the certificates for my routes?
Looks like there is some disccusion of this over in .
users mailing list
users lists openshift redhat com