F5 TMM pods not starting on worker node

Issue

• We have an application that runs on nodes which have a specific mcp applied which sets up specific kernel args:

spec:
   additionalKernelArgs:
   - nosmt
   cpu:
      isolated: 2-13
      reserved: 0-1
   hugepages:
      defaultHugepagesSize: 2M
      pages:
      - count: 58000
        node: 0
        size: 2M

• what we are seeing is that it works fine on one node, but not both.
• If we reboot the one that works, it seem to start working on the other node. This happens each time we try to run TMM with two replicas.
• Currently worker 1 is working and worker-0 is not.