This message was deleted.
# rke2
a
This message was deleted.
a
maybe worth noting is that we haven't seen this in a staging or a test cluster with the same configurations - although they see far lower rates of change and aren't anywhere as large
also, of interest, the faulty cilium instance that needs a restart can't run commands like
cilium nodeid list
, it sits forever - restart the pod, and the command returns instantly with results
https://github.com/cilium/cilium/issues/24574 appears similar to what we see
In case anyone else comes across this, upgrading to Cilium 1.13.4 fixed the issue with the wireguard deadlock