Public Cloud Network Services - FAQ

Load Balancer is offered in different sizes (S/M/L) to best fit our customer needs. These different sizes are defined via flavors. As of today, to change the size of your Load Balancer, you will need to spawn a new one, configure it the same way (with the same backends as the old one), and reconnect the Floating IP to the new one. Then the old Load Balancer can be deleted.

Yes, provided you configure the network connectivity between the Load Balancer and your bare metal server (either through the private network (vRack) or through a public IP). If your network is configured correctly, a Load Balancer can redirect traffic to members in different Public Cloud regions.

Octavia Load Balancer is not currently supported with MKS. This is currently on the roadmap.

Yes, this feature has been available since we rolled out Zed version in June 2023. You can find more information here.

Yes, we offer Active/Standby mode for all plans S/M/L.

No. A new type of IP has been introduced for this purpose, Floating IP. It's more flexible and works on a pay-as-you-go model for automation purposes. With Floating IP, you can use a single Floating IP per Load Balancer service.

Yes, multiple listeners (frontends) and backends can be configured. There is a limitation of using a single Floating IP per Load Balancer.

First of all, the values shown are only a rough estimate of the Load Balancer's capabilities. The price will not increase. Pricing is linked to the flavors (small, medium, large) and we cannot change the flavor at this time. This is up to customers' orchestration services.

It is up to the customer to monitor the load balancer using the metrics feature and to change the flavor accordingly.

In a Public to Public architecture, the outbound traffic is managed by the Gateway component. If you use that architecture, you should size the Gateway accordingly.

Gateway is the product name for the Distributed Virtual Router (DvR) component in OpenStack. It is offered with High Availability and Service Level Agreement.

Gateway is offered in different sizes (S/M/L) to best fit our customer needs. These different sizes are defined via QoS policies. As of today, to change the size of Gateway, a change of QoS policy between S/M/L can be done.

No, a Gateway can only be used in one region.

If you have a private network that spreads over multiple regions (thanks to the same VLAN Id), you need to spawn a Gateway in each region.

For example, the following architecture can be used:

Yes, that's exactly the use case for Gateway (L3 router with SNAT option). Currently, only instances in private mode of networking that are connected to a single-region private network are supported. For more information, please refer to our Concepts - Public Cloud Networking guide.

Yes, you can use an L3 router without the SNAT option through the OpenStack GUI / CLI / Terraform. In that case, the bandwidth limits are driven by the quality of service on the instance private bandwidth. Hence, choosing an S flavor would not impact the performances.

No, inter-region routing is not supported.

It depends on the usage:

• For outbound use cases we offer a public IP which is included in the Gateway price. This IP is associated to the instantiated Gateway and cannot be moved from one gateway to another. In other words, the IP used for outbound traffic is not a 'Floating IP'. The only use case where this IP is kept is when the gateway size is changed (e.g. from S to M). If this is creating friction for your use case, please upvote for this roadmap item.
• For inbound use cases (to expose a service running on a private instance to the Internet) you need to have a Floating IP to attach via Gateway to your instance or network service.

You can create a private network in a selected region and attach Gateway to it. Then, while creating an instance, select "Attach to private network" and confirm it using the button "Yes, I want my instance to be fully private".

Yes, the OHVcloud Anti-DDoS infrastructure is activated on all those IPs. You can view that information in the Bare Metal Cloud > Network > IP section of the Control Panel.

Two options can be used:

• Using a "jump host" (SSH proxy): You need to use another instance that has Floating IP (allowing for external access) and a private port in the same private network that your new instance is in. Log on to it and connect via SSH to the private IP of your new instance.
• Attaching a Floating IP address (at least for the time of maintenance) to your newly created instance and connect to it using that Floating IP.

A Floating IP is a flexible public IP address that can be associated to a private port only (port on a private network in vRack). You will need also a Gateway for it to work.

The pool for a Floating IP must be "Ext-Net" and you can associate it to a port on the private network.

Yes, the OHVcloud Anti-DDoS infrastructure is activated on all the Floating IPs. You can view that information in the Bare Metal Cloud > Network > IP section of the Control Panel.