...
- 2 servers. 2 AIO servers with HA (high availability), the controller functionality and storage functionality will be deployed at the 2 servers with standby-active mode managed by "service management". If one server or one service in one server has error, it will be switched from active to standby one to maintain the service availability.
- 2 dedicated controller servers within the storage inside, and multiple worker nodes.
- 2 dedicated controllers servers, 2/3 (more) dedicated storage servers, and multiple worker nodes. This is the standard configuration.
- Distributed Cloud architecture support.
...
In the Bronze release, the INF deliverable started to work with ODULOW team, the ODULOW deliverable asked the following requirement:
K8S features and plugins | Required for OSC lab (Y/N) | Comments |
Multus | Y | |
Calico | Y | Easy to deploy |
Native Huge Pages | Y | Best to have numa aware support but k8s don’t support yet |
SR-IOV Net Device Plugin | Y | Use for NIC SRIOV management. Better to extend support of VC sriov. |
Node Feature Discovery | Y | Easy to deploy |
FPGA device plugin | Y | Manager PAC N3000 VF devices in k8s |
K8s CPU manager/topology manager | Y | Key feature for CPU isolation |
In the Cherry release, will keep support those features, and will start to do the interoperability with the following components:
...