High Availability is a key issue for Telco NFV cloud. The current Yardstick test cases include HA cases, which mostly focusing on component and OpenStack-service HA. When we go to field deployment, the resource pool could be more than hundreds. How to ensure high availability and reliability for such large scale resource pool is a key issue for Telco Operators. Therefore HA testing for large scale resource pool is necessary to benchmark the reliability scale for the pool. Test cases include but not limited to:
1) Capability of setup/delete/move large number of VM in a limited time.
2) Capability of moving a certain number of VM to the spare resource pool, when large number of servers are shut down.
3) Capability of moving a certain number of VM when operation/service/storage network fails
4) Capability of figuring out warning from hardware and evacuate VM in advance
We will bring real Telco failure usecase into the project as reference so as to define suitable test spec, and work on test cases for large scale resource pool HA.
high availability for OPNFV
- Telco HA test cases for large scale resource pool
Desired project timeline/completion date:
Usecase analysis and test spec definition
Test Case Development
Mentor(s) & contact info:
I have accomplished the Reliability testing which includes the Large-scale hypervisor node failure detection and the Control nodes failover test. Then I accomplished the case of Hypersior service plane single network port failure (VIM), the case of Hypersior service plane all network port failure (VIM), the case of virtual machine update test and the case of Storage QoS test. I have attached a link to GITHUB and some screenshots of cases.