calico/node is not ready: BIRD is not ready: Error querying BIRD: unable to connect to BIRDv4 socket

[l1douhua]

The calico-node of a node in the cluster is not Running. I restarted the calico-node of the node and found an error, [calico/node is not ready: BIRD is not ready: Error querying BIRD: unable to connect to BIRDv4 socket]. I thought it was because of the network card problem, so I restarted the node and it recovered. Why does this problem occur when the network card has not been changed?

Below are the cluster versions I use:

kubernetes:v1.28.2
containerd:1.6.33
docker.io/calico/apiserver:v3.26.3
docker.io/calico/kube-controllers:v3.26.3
docker.io/calico/node:v3.26.3
docker.io/calico/typha:v3.26.3

[caseydavenport]

We're going to need more information to help with this one - are there any logs from the calico/node pods for example?

[l1douhua]

No exception logs are saved. Can you tell from the errors provided where the problem occurred? @caseydavenport

[caseydavenport]

Error querying BIRD: unable to connect to BIRDv4 socket

This error is very generic, and generally just means that BIRD failed to start. There's no way to really tell without the logs from the calico/node pod that was encountering this issue.