Description of problem: When name one of the interfaces on the host matches the scientific notation (e.g. 5e10), it causes a panic in knmstate. This was supposed to be solved by https://bugzilla.redhat.com/show_bug.cgi?id=1926986, but seems to be still happening. Reported in https://bugzilla.redhat.com/show_bug.cgi?id=1926986#c15. """ device name was 94475496822e234 and we had to kill the associated pod (sandboxid) and after restarting the nmstate-handler, we do not have anymore the error : created by sigs.k8s.io/controller-runtime/pkg/internal/controller.(*Controller).Start.func1 /remote-source/app/vendor/sigs.k8s.io/controller-runtime/pkg/internal/controller/controller.go:212 +0x45e panic: interface conversion: interface {} is float64, not string [recovered] panic: interface conversion: interface {} is float64, not string """ Version-Release number of selected component (if applicable): OCP 4.8.27 How reproducible: Sometimes. This only reproduces when the veth interface matches the notation. Steps to reproduce without relying on the number name: 1. Create a veth pair where the host-side interface's name matches the scientific notation Actual results: knmstate's pod panics Expected results: knmstate reports the interface with its name and does not crash. Additional info:
This was fixed on the standalone operator. knmstate is not shipped with CNV in 4.11. Closing.