Question: How to fix when Spinnaker containers gone Crashloopbackoff?


Answer: If you notice below exceptions

  • Error: Could not create the Java Virtual Machine.
  • Error: A fatal exception has occured. Program will exit.

This is mostly because the Spinnaker containers are not allocating enough resources from the Cluster. Note that worker nodes in the cluster need to have a minimum of 8GB RAM for smooth operations. 

To manually increase  resources, edit each and every servers and update the JVM parameters shown below as an example:

spec:
  affinity: {}
  containers:
  - env:
    - name: JAVA_OPTS
      value: -Xmx300m 
            --XX:MaxRAM=600m -Xmx500m
  • Also, most of the cases this issue is observed in users who are using halyard version prior to 1.29. By updating the halyard version to 1.29 or later no manual changes required for allocating resources to Spinnaker services.