While enabling the Microservices Infrastructure (CMSP) on Nutanix Prism Central, the deployment failed during the post-deployment prechecks with an error indicating that Prism Central was not compatible with enabling Microservices Infrastructure.
* Root Cause Analysis:
- When CMSP is enabled, Prism Central deploys an internal Kubernetes-based platform to host Nutanix microservices.
- For CMSP deployment on Prism Central, the internal CMSP services use the following default networks:
- 192.168.5.0/24 – Kubernetes infrastructure network
- 10.100.0.0/16 – Kubernetes Pod network
- 10.200.32.0/24 – Kubernetes Services network
- The upstream network environment must not overlap with these network ranges.
- Since Kubernetes cannot differentiate between internal container traffic and the production management network, CMSP prechecks intentionally block the deployment to prevent potential routing conflicts.
* Solution:
- If the environment requires using these subnet ranges in the upstream network, you must change the CMSP internal network ranges before deploying Prism Central.
- Since the deployment had already failed, the first step was to clean up the incomplete CMSP deployment and then configure custom Pod and Service CIDRs.
1. SSH into PCVM and clean up the failed deployment
python /home/docker/msp_controller/bootstrap/msp_tools/cmsp-scripts/cmsp-cleanup.py --force
2. Configure new CMSP networks
Edit the Genesis gflags: ~/serviceability/bin/edit-aos-gflags --service=genesis
Update the default networks to unused subnets. For example:
--cmsp_pod_network=172.16.0.0/16
--cmsp_service_ip_network=172.17.32.0/24
3. Restart Genesis
genesis restart
4. Re-deploy CMSP
Navigate to the CMSP scripts directory:
cd /home/docker/msp_controller/bootstrap/msp_tools/cmsp-scripts/
Then execute: python cmsp-enable.py
5. Monitor the deployment progress
Option 1: Check infrastructure tasks
ecli task.list component_list=infra
Option 2: Monitor the CMSP deployment pipeline
mspctl task list
