0

calico pod on master node is in CrashLoopBackOff because of "unable to connect to BIRDv4 socket"

    lab@master:~$ kubectl -nkube-system get po -o wide
    NAME                                       READY   STATUS             RESTARTS         AGE     IP               NODE       NOMINATED NODE   READINESS GATES
    calico-kube-controllers-74d5f9d7bb-k44d2   1/1     Running            19 (14m ago)     22d     10.10.219.106    master     <none>           <none>
    calico-node-bfqsb                          1/1     Running            2 (5h36m ago)    5h47m   192.168.56.111   worker01   <none>           <none>
    calico-node-trc9k                          0/1     CrashLoopBackOff   6 (117s ago)     8m35s   192.168.56.110   master     <none>           <none>
    coredns-5dd5756b68-4xvp6                   1/1     Running            17 (14m ago)     22d     10.10.219.107    master     <none>           <none>
...

The way to go is this: https://github.com/projectcalico/calico/issues/2834

However, the document does not specify WHERE (in which file) to set/add one of these options. Does anyone know?

1 Answer 1

0

Reason: There are too many NICs and calico is choosing wrong one.

To help Calico to choose desired NIC / IP address (that will be used to route between nodes) add IP auto-detection in calico-node DAEMONSET manifest.
Somehow u should understand that out of the configuration documentation:

Todo:

1. Backup current daemonset/calico-node
kubectl -nkube-system get daemonset calico-node -o yaml > calico-node.DAEMONSET.orig

2. Add IP autodetection in  calico-node DAEMONSET manifest:
Note: I have choosen IP_AUTODETECTION_METHOD=can-reach=192.168.56.110 because it works the best for my case. U might want to try other options: https://docs.tigera.io/archive/v3.8/reference/node/configuration#ip-setting

kubectl -nkube-system edit daemonset/calico-node

Add it in this block:
        - name: CLUSTER_TYPE
          value: k8s,bgp
        - name: IP
          value: autodetect
        - name: CALICO_IPV4POOL_IPIP
          value: Always
        - name: CALICO_IPV4POOL_VXLAN
          value: Never
        - name: CALICO_IPV6POOL_VXLAN
          value: Never
        - name: IP_AUTODETECTION_METHOD   <--- add this
          value: can-reach=192.168.56.110 <--- add this
        - name: FELIX_IPINIPMTU

3. calico-node pods will be restarted with new config automatically. 
kubectl -nkube-system get po -l  k8s-app=calico-node -o wide
NAME                READY   STATUS    RESTARTS   AGE   IP               NODE       NOMINATED NODE   READINESS GATES
calico-node-dj2cl   1/1     Running   0          32m   192.168.56.110   master     <none>           <none>
calico-node-lq4h5   1/1     Running   0          32m   192.168.56.111   worker01   <none>           <none>

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .