I have started seeing this since RKE2 1.31... on rhel 9.4 and now also on 9.6 rke2-ingress-nginx-controller takes a long time to terminate (over 10-min) have noticed this during maintenance when I am rebooting 1 node at a time (3-node cluster) anyone else have noticed this?

There was a period of 1-2 months of release where the underlying BCI Base image removed

procps

package as default and thus no

pkill

was available to nginx to cleanly terminate. That got fixed with April 2025+ releases of RKE2 (so for 1.31.8+rke2r1 and newer)

Thanks! so this fix should be in rke2 1.32.5 then?

Yep