Tenant network always loss ipvlan link

[antzjm]

Is this a BUG REPORT or FEATURE REQUEST?:

Uncomment only one, leave it on its own line:

bug

What happened: After I restart host infra, The tenant network with ipvlan always loss links. And I restart network-watch daemon-set to recover. What you expected to happen: Restart host and keep fine. How to reproduce it:

Anything else we need to know?:

Environment:

• DANM version (use 4.1.0-2):

• Kubernetes version (use v1.16.2):

• DANM configuration (K8s manifests, kubeconfig files, CNI config file):

• tenant.yaml

• apiVersion: danm.k8s.io/v1 kind: TenantNetwork metadata: annotations: kubectl.kubernetes.io/last-applied-configuration: | {"apiVersion":"danm.k8s.io/v1","kind":"TenantNetwork","metadata":{"annotations":{},"name":"internal","namespace":"cran1"},"spec":{"NetworkID":"internal","NetworkType":"ipvlan","Options":{"allocation_pool":{"end":"192.168.1.254","start":"192.168.1.2"},"cidr":"192.168.1.0/24","container_prefix":"internal","host_device":"ens1f1"}}} creationTimestamp: "2021-03-31T05:50:44Z" generation: 3932 name: internal namespace: cran1 resourceVersion: "18383421" selfLink: /apis/danm.k8s.io/v1/namespaces/cran1/tenantnetworks/internal uid: df1526ba-95c6-45c1-a8a1-a214f9306ae9 spec: NetworkID: internal NetworkType: ipvlan Options: alloc: vIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAE= allocation_pool: end: 192.168.1.254 start: 192.168.1.2 cidr: 192.168.1.0/24 container_prefix: internal host_device: ens1f1 vlan: 1711

• OS (e.g. from /etc/os-release): CentOS Linux release 7.8.2003 (Core)

• Kernel (e.g. uname -a): Linux controller-1 4.19.162-1.el7.centos.ncir.x86_64

• Others:

[Levovar]

@antzjm : just to check if I understand the problem description 1: you deploy above network -> netwatcher creates the host interface as usual 2: you restart the host -> host VLAN interface disappears 3: you restart netwatcher on this host -> host VLAN interface is re-created again is this correct?

[antzjm]

@Levovar Yes. When host VLAN interface loss, pod can not created due to CNI network plugin fail.

[Levovar]

# ip l | grep 2000
# kubectl create -f test_net.yaml
clusternetwork.danm.k8s.io/external created
# ip l | grep 2000
117: external.2000@infra-bond: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP mode DEFAULT group default
# reboot
...
# ip l | grep 2000
91: external.2000@infra-bond: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP mode DEFAULT group default

I'm not sure what's your issue exactly, but the described scenario works as intended

[Levovar]

closing as the reported scenario cannot be reproduced, and is working as intended