Multicast traffic to VMs fails when using active-backup bond
Issue
- We have a KVM hypervisor that runs a guest that receives multicast. The virtual network is connected to a bridge on the hypervisor to a bonded network interface in active-backup mode.
- In case of a breakdown of the current active path of the bond, it requires the IGMP router query to be received before the multicast reception path to the guest is restored. The multicast querier is configured to poll once every minute; this results in an outage of multicast reception for upto 1 minute on failover.
- Multicast traffic to VMs fails when using active-backup bond
Environment
- Red Hat Enterprise Linux, all versions
- KVM hypervisor
- Bridge used to provide network connectivity to VMs
- Physical interfaces in active-backup bond:
switch --- ethX/ethY --- bond0 --- br0 --- vnetX --- VM ethX - Guests subscribe to multicast groups
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
