Upgrade 1.28.8 -> 1.28.9 left zombie process in kube-proxy pod

[xhejtman]

   7087 ?        Sl     0:23 /var/lib/rancher/rke2/data/v1.28.9-rke2r1-fc5244538cf2/bin/containerd-shim-runc-v2 -namespace k8s.io -id d0542cb548d6c818d8a
   7107 ?        Ss     0:00  \_ /pause
   7139 ?        Ssl    0:41  \_ kube-proxy --cluster-cidr=10.42.0.0/16 --conntrack-max-per-core=0 --conntrack-tcp-timeout-close-wait=0s --conntrack-tcp-
   7173 ?        Zs     0:00      \_ [timeout] <defunct>

is it something to worry about?

[brandond]

I don't think so, are there any errors in the pod logs or other things that appear to be functioning properly?

[xhejtman]

cat 0.log
2024-05-02T22:24:01.367575659+02:00 stderr F I0502 20:24:01.367425       1 node.go:141] Successfully retrieved node IP: 10.16.48.12
2024-05-02T22:24:01.398930705+02:00 stderr F I0502 20:24:01.398852       1 server.go:632] "kube-proxy running in dual-stack mode" primary ipFamily="IPv4"
2024-05-02T22:24:01.400040681+02:00 stderr F I0502 20:24:01.399989       1 server_others.go:152] "Using iptables Proxier"
2024-05-02T22:24:01.400069232+02:00 stderr F I0502 20:24:01.400047       1 server_others.go:421] "Detect-local-mode set to ClusterCIDR, but no cluster CIDR for family" ipFamily="IPv6"
2024-05-02T22:24:01.400075168+02:00 stderr F I0502 20:24:01.400067       1 server_others.go:438] "Defaulting to no-op detect-local"
2024-05-02T22:24:01.400126655+02:00 stderr F I0502 20:24:01.400106       1 proxier.go:250] "Setting route_localnet=1 to allow node-ports on localhost; to change this either disable iptables.localhostNodePorts (--iptables-localhost-nodeports) or set nodePortAddresses (--nodeport-addresses) to filter loopback addresses"
2024-05-02T22:24:01.400314863+02:00 stderr F I0502 20:24:01.400282       1 server.go:846] "Version info" version="v1.28.9+rke2r1"
2024-05-02T22:24:01.400485706+02:00 stderr F I0502 20:24:01.400451       1 server.go:848] "Golang settings" GOGC="" GOMAXPROCS="" GOTRACEBACK=""
2024-05-02T22:24:01.40119637+02:00 stderr F I0502 20:24:01.401144       1 config.go:97] "Starting endpoint slice config controller"
2024-05-02T22:24:01.401204574+02:00 stderr F I0502 20:24:01.401169       1 shared_informer.go:311] Waiting for caches to sync for endpoint slice config
2024-05-02T22:24:01.401338236+02:00 stderr F I0502 20:24:01.401300       1 config.go:188] "Starting service config controller"
2024-05-02T22:24:01.401343354+02:00 stderr F I0502 20:24:01.401309       1 shared_informer.go:311] Waiting for caches to sync for service config
2024-05-02T22:24:01.401839729+02:00 stderr F I0502 20:24:01.401794       1 config.go:315] "Starting node config controller"
2024-05-02T22:24:01.401848416+02:00 stderr F I0502 20:24:01.401803       1 shared_informer.go:311] Waiting for caches to sync for node config
2024-05-02T22:24:01.501458938+02:00 stderr F I0502 20:24:01.501353       1 shared_informer.go:318] Caches are synced for endpoint slice config
2024-05-02T22:24:01.501514906+02:00 stderr F I0502 20:24:01.501436       1 shared_informer.go:318] Caches are synced for service config
2024-05-02T22:24:01.501992204+02:00 stderr F I0502 20:24:01.501943       1 shared_informer.go:318] Caches are synced for node config

this is all I have in the kube-proxy logs.

[brandond]

seems fine then? defunct processes aren't a problem per-se, it just indicates that kube-proxy ran a subcommand and didn't reap the pid when it exited.

[xhejtman]

yes, just didn't happen in 1.28.8, happens on 1.28.9 on all three nodes. Just letting you know ;)

[ffraz]

kube-proxy versions without timeout zombie:

• v1.27.11+rke2r1
• v1.28.8+rke2r1
• v1.29.3+rke2r1
• v1.29.13+rke2r1 (Pre-release)
• v1.30.9+rke2r1 (Pre-release)
• v1.31.5+rke2r1 (Pre-release)