You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Setup: Kubernetes on AWS EC2-instance running. 1 master and 2 nodes.
CNI plugin used : Multus with amazon-vpc-cni-k8s configured to be used for secondary interface on the pod.
Version: latest Multus code + latest amazon-vpc-cni-k8s code.
When Multus plugin is used, the pod comes up with multiple interfaces. Ideally when we want to run pods in k8's hosted in AWS, we would want to use amazon-vpc-cni-k8s cni to provide a secondary interface to the pod.
Using amazon-vpc-cni-k8s plugin as secondary cni plugin, plugin errors out with the below message:
2018-10-11T11:46:42Z [INFO] Starting CNI Plugin v1.2.0-6-g645535ed-dirty ...
2018-10-11T11:46:42Z [INFO] Received CNI add request: ContainerID(c3204c7ddb2a1a094e5275018d42c745539f026649fbb55b1296ce6a2111986f) Netns(/proc/130265/ns/net) IfName(net1) Args(IgnoreUnknown=1;K8S_POD_NAMESPACE=default;K8S_POD_NAME=samplepod4;K8S_POD_INFRA_CONTAINER_ID=c3204c7ddb2a1a094e5275018d42c745539f026649fbb55b1296ce6a2111986f) Path(/opt/cni/bin) argsStdinData({"cniVersion":"0.3.0","name":"aws-cni","type":"aws-cni","vethPrefix":"eni"})
2018-10-11T11:46:42Z [INFO] Received add network response for pod samplepod4 namespace default container c3204c7ddb2a1a094e5275018d42c745539f026649fbb55b1296ce6a2111986f: 172.31.19.103, table 0
2018-10-11T11:46:42Z [ERROR] Failed to setup NS network setup NS network: failed to add default route: file exists
2018-10-11T11:46:42Z [ERROR] Failed SetupPodNetwork for pod samplepod4 namespace default container c3204c7ddb2a1a094e5275018d42c745539f026649fbb55b1296ce6a2111986f: setup NS network: failed to setup NS network: setup NS network: failed to add default route: file exists
This is because amazon-vpc-cni-k8s is trying to setup default route but it's already been set beforehand by the primary Multus plugin. So the plugin errors out.
CNI's are supposed to be resilient if the route already exists but in case of amazon-vpc-cni-k8s it fails.
Request to allow amazon-vpc-cni-k8s plugin to work even if the default route is already set.
There are no errors in L-IPAMD logs.
The text was updated successfully, but these errors were encountered:
Deepthidharwar
changed the title
amazon-vpc-cni-k8s cni erros out due to setupNS network failing as 'file exists' when default route is already established.
amazon-vpc-cni-k8s cni errors out due to setupNS network failing as 'file exists' when default route is already established.
Oct 17, 2018
Setup: Kubernetes on AWS EC2-instance running. 1 master and 2 nodes.
CNI plugin used : Multus with amazon-vpc-cni-k8s configured to be used for secondary interface on the pod.
Version: latest Multus code + latest amazon-vpc-cni-k8s code.
When Multus plugin is used, the pod comes up with multiple interfaces. Ideally when we want to run pods in k8's hosted in AWS, we would want to use amazon-vpc-cni-k8s cni to provide a secondary interface to the pod.
Using amazon-vpc-cni-k8s plugin as secondary cni plugin, plugin errors out with the below message:
This is because amazon-vpc-cni-k8s is trying to setup default route but it's already been set beforehand by the primary Multus plugin. So the plugin errors out.
CNI's are supposed to be resilient if the route already exists but in case of amazon-vpc-cni-k8s it fails.
Request to allow amazon-vpc-cni-k8s plugin to work even if the default route is already set.
There are no errors in L-IPAMD logs.
The text was updated successfully, but these errors were encountered: