Calico 3.31.2 - whisker CrashLoopBackOff

[kkbruce]

Our operating system defaults to disabling IPv6, and we are not sure if this is causing the whisker container to fail to start correctly.

$ kubectl get pod -n calico-system
NAME                                       READY   STATUS             RESTARTS           AGE
calico-apiserver-868cfdf9fb-6d8kb          1/1     Running            0                  12d
calico-apiserver-868cfdf9fb-kvwjp          1/1     Running            0                  12d
calico-kube-controllers-58cb95797f-6jnzh   1/1     Running            0                  12d
calico-node-4nclv                          1/1     Running            0                  12d
calico-node-7xk9n                          1/1     Running            0                  12d
calico-node-mgd9s                          1/1     Running            0                  12d
calico-typha-776d6db587-7rptf              1/1     Running            0                  12d
calico-typha-776d6db587-sb7dw              1/1     Running            0                  12d
csi-node-driver-8rz6x                      2/2     Running            0                  12d
csi-node-driver-kqmj6                      2/2     Running            0                  12d
csi-node-driver-mjzzb                      2/2     Running            0                  12d
goldmane-65dcd4f69b-v4lrs                  1/1     Running            0                  12d
whisker-5d6c49988-jjsvw                    1/2     CrashLoopBackOff   3626 (2m33s ago)   12d


$ kubectl logs whisker-5d6c49988-jjsvw -n calico-system
Defaulted container "whisker" out of: whisker, whisker-backend
2025/12/14 09:38:38 [emerg] 1#1: socket() [::]:8081 failed (97: Address family not supported by protocol)
nginx: [emerg] socket() [::]:8081 failed (97: Address family not supported by protocol)


$ calicoctl version
Client Version:    v3.31.2
Git commit:        dd5575465ed825f86a3bc10f17bb5f7dc106d0f0
Cluster Version:   v3.31.2
Cluster Type:      typha,kdd,k8s,operator,bgp,kubeadm

[coutinhop]

@kkbruce can you share some more details about your setup? if possible, share your cluster's FelixConfiguration and IPPools, do you have IPv6 enabled in Calico but not in the OS in the underlying nodes?

[caseydavenport]

I think this is likely an IP family issue - looks similar to https://github.com/projectcalico/calico/issues/10956, which was fixed by https://github.com/tigera/operator/pull/4204

[caseydavenport]

Not yet released - should be in the next v3.31 patch release.

[kkbruce]

Hi @caseydavenport Is there a planned release date for the next version?