Test the following scenario :

• Install 6 nodes cluster with OCS block storage
• Verify VMS with io='native'
• Create and start 1500 small Cirros VMS on 6 worker nodes.
• Create and start 6 windows VMS
• Create multiple namespaces with 15 NS, 100 VMS in each one
• Create the following resource quotas on the Cirros VMS: 
• requests.memory 64M
• requests.cpu 100Mi
• requests.storage 1Gi

• Create the following resource quotas on the Windows VMS: 
• requests.memory 1Gi
• requests.cpu 1
• requests.storage 16Gi

• Create workload inside 100 VMS
• Create workload OTO 6 worker nodes VMS
• Check all VMS are alive (SSH connection)
• Run continually mass migration for 12 hours with MCP rollout as in BUG 2207682
• Record system metric’s and export it to the performance team Grafana server
• Check general exit criteria passed successfully : 

• No critical alerts at the Prometheus during the test

• VM scheduling is well-balanced across the nodes

• No pod crash/restarts

• System resource utilization is within range and no degradation was noticed

• All VMS stay alive during the migrations, no crashed or blue screen was notice

• Storage system state is healthy

• For all live migrations tests : 
• All vmim objects were created
• Status of vmim was finished successfully - Phase: Succeeded
• VMI Migration State was finished successfully with no error and end timestamp
• VMI node changed
• VMI age stayed the same (live migrated)
• Check all VMS are alive (SSH connection) and uptime indicated the VM was not rebooted

Full test plan :

https://docs.google.com/document/d/1dK5XPgU8fooMONmMGIlAh_LWAySBHue_wMo63XOwIO0/edit?usp=sharing