I am facing following issue when creating pod using sriov network. when i see output of device driver using
$**dmesg**
*mlx5_core 0000:b5:00.0: mlx5_cmd_check:772:(pid 5271): CREATE_SQ(0x904) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0xd61c0b)*
seems some issue in communication with hardware. below listed system details:
PCI: b5:00.0 Ethernet controller: Mellanox Technologies MT28841
Driver: mlx5_core (5.0-0.0)
Card: ConnectX-6 Dx
OS version: Ubuntu 18.04.6 LTS
Kernel: Linux 5.4.0-89-generic
Architecture: x86-64
pod yaml:
apiVersion: v1
kind: Pod
metadata:
name: mellonox-test-pod
annotations:
k8s.v1.cni.cncf.io/networks: sriov-rdma-net1
spec:
containers:
- name: mellonox-alpine
image: alpine
imagePullPolicy: Never
command: [ "/bin/bash", "-c", "--" ]
args: [ "while true; do sleep 300000; done;" ]
resources:
requests:
mellanox.com/mlnx_sriov_rdma: '1'
limits:
mellanox.com/mlnx_sriov_rdma: '1'
pod creation result:
$kubectl describe pod mellonox-test-pod
`"error bringing interface up in container ns: \"invalid argument\""`
After that when i see device driver output, got this issue *mlx5_core 0000:b5:00.0: mlx5_cmd_check:772:(pid 5271): CREATE_SQ(0x904) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0xd61c0b)*.
Would be great to receive any help. Thanks in advance.
can you maybe try to be more specific in the network annotations? seems something related to the namespace, add metadata.namespace=target_namespace.
Be aware the target_namespace needs to be the same in both the Pod spec and the SriovNetowrk CR.
Cheers.