Enabling VM communication - Part 1

Scenario

The virtual machines exercise10-a and exercise10-b are operational and with their IP addresses correctly configured. However, they are not communicating with each other as expected.

Your task is to identify and troubleshoot the issue preventing communication between the VMs.
Resolve it to restore connectivity.

The steps to fix exercise10 are:

Console

{openshift_console_url}

Username

userx

Password

{password}

Go to virtualization → Virtual Machines - select project for the assigned user account

Click on exercise10-a virtual machine, Configuration and Network, verify it’s using exercise10-nad as secondary network

Click on exercise10-b virtual machine, Configuration and Network, verify it’s using exercise10-nad as secondary network

On left side menu click Networking → NetworkAttachmentDefinition

Click exercise10-nad and check it’s using Linux Bridge

O left side menu click Networking → NodeNetworkConfigurationPolicy and verify that it is empty

O left side menu click Networking → NodeNetworkConfiguration -→ click on the list button on the top right, and check there’s a linux bridge interface on nodes

Click linux bridge → br-userx

Verify that there are only veth interfaces. It means that there is no physical network attached on this linux bridge

On left side menu click Networking → NetworkAttachmentDefinitions and Create Network Attachment Definition button.

Type exercise10-nad-layer2 as name and select OVN Kubernetes L2 overlay network on Network Type and click on Create button

On left side menu click Virtualization → VirtualMachines and stop virtual machine exercise10-a

On left side menu click Virtualization → VirtualMachines and stop virtual machine exercise10-b

Click on exercise10-a virtual machine, Configuration and Network,

Edit the virtual machine network

Change Network from exercise10-nad to exercise10-nad-layer2

Start the virtual machine exercise10-a

Repeat steps 15, 16, 17, and 18 for the exercise10-b virtual machine

Run grade to validate the exercise

lab grade exercise10

Command line (CLI)

OpenShift login command

{login_command}

Go to the assigned namespace-userx

oc project namespace-userx

List the virtual machines

oc get virtualmachine

Check the virtual machines are running on different nodes

oc get pods -o wide

Check communication is not working

virtctl ssh lab-user@exercise10-a \
-c 'ping -c 5 192.168.0.11'

Get the secondary network used on vms a and b

oc get vmi exercise10-a -n namespace-userx \
-o jsonpath='{.spec.networks[?(@.multus)].multus.networkName}'

oc get vmi exercise10-b -n namespace-userx \
-o jsonpath='{.spec.networks[?(@.multus)].multus.networkName}'

Get the secondary network yaml

oc get network-attachment-definitions.k8s.cni.cncf.io exercise10-nad -o yaml

Ensure that the NaD (NetworkAttachmentDefinition) is configured to use a br-userx interface. Verify that this Linux bridge is not attached to any physical port (one or more veth* port are used), meaning virtual machines will only be able to communicate if they are running on the same node. Therefore, you must recreate the NaD to use layer2 networking instead of Linux bridge. Create layer2-nad.yaml and apply it.

oc get nns -o yaml | grep -A 2 -B 73 br-userx | grep -A4 port

vim layer2-nad.yaml

apiVersion: k8s.cni.cncf.io/v1
kind: NetworkAttachmentDefinition
metadata:
  name: exercise10-nad-layer2
  namespace: namespace-userx
spec:
  config: '{"name":"exercise10-nad-layer2","type":"ovn-k8s-cni-overlay","cniVersion":"0.3.1","topology":"layer2","netAttachDefName":"namespace-userx/exercise10-nad-layer2"}'

oc apply -f layer2-nad.yaml

shutdown the virtual machines exercise10-a and exercise10-b

virtctl stop exercise10-a
virtctl stop exercise10-b

change the virtual machines secondary network

oc patch vm exercise10-a -n namespace-userx --type='json' \
-p='[{"op": "replace", "path": "/spec/template/spec/networks/1/multus/networkName", "value": "exercise10-nad-layer2"}]'

oc patch vm exercise10-b -n namespace-userx --type='json' \
-p='[{"op": "replace", "path": "/spec/template/spec/networks/1/multus/networkName", "value": "exercise10-nad-layer2"}]'

11.Start the virtual machines

virtctl start exercise10-a
virtctl start exercise10-b

After virtual machines starts wait some time and check the communication again

virtctl ssh lab-user@exercise10-a \
-c 'ping -c 3 192.168.0.11'

What you learned

In this exercise, you learned that to enable communication between virtual machines, a secondary network using OVN layer2 must be configured instead of a Linux Bridge. The OVN localnet and Linux Bridge options are not viable because no NodeNetworkConfigurationPolicy has been defined, which means the bridge does not have a physical port attached.