This article addresses an upgrade failure in VCF Operations for Networks (formerly Aria Operations for Networks/vRNI) where the FlinkTaskManager Java services fail to start or stay running during the upgrade process.
Symptoms: The upgrade process stalls or rolls back. Inspection of services reveals that flink-taskmanager is not in a running state, preventing the cluster from reaching a healthy status for the update.
VCF Operations for Networks 9.x
Configurations: Distributed deployments or environments with strict network security policies/firewalls.
The failure is caused by a connectivity blockage. The upgrade requires specific communication ports to be open for Flink services to coordinate tasks. If these ports are not reachable or if the virtual machines are subject to restrictive security policies (such as NSX Distributed Firewall rules without proper exclusions), the FlinkTaskManager service will fail to initialize.
This is caused due customer have not enabled the Ports, 8030 ,8088 , 8025 , 45454(No details in the Doc).
Firewall ports should be enabled as per Ports.broadcom.com for connectivity
Follow these steps to resolve the connectivity issues and allow the upgrade to proceed:
bashservice flink-taskmanager status
bashservice flink-taskmanager restart
Ports.broadcom.com/home/VMware-Aria-Operations-for-Networks
Documentation Deficiency identified regarding specific port updates for version 9.0.x.