Symptoms
The VMware SD-WAN Spoke Edges assigned to the SD-WAN Hub Cluster member fails to receive the endpoint information of the Hub Cluster member, leading to no overlays between the Spoke Edges and the Hub Clusters.
Purpose
This KB article documents the issue encountered under internal ticket #74316.
The behavior is seen with a customer using a Edge Cluster Hub and Spoke topology.
Cause
The sequence of events that triggers this issue is as follows:
A Spoke Edge is connected to a Hub Edge that is a member of a Cluster.
The tunnel between this Hub Edge and the VMware SD-WAN Super Gateway goes down but the tunnel between the Spoke Edge and the Hub Edge remains intact.
As expected, the Super Gateway will move the Spoke Edge to another member of the Cluster, but the Spoke Edge will reject the update since the tunnel to its Hub Edge is up and running.
The Spoke Edge will send a cluster assignment message to the Super Gateway using the ID for the Hub Edge it is connected to, the Super Gateway will create a new cluster info entry with no tunnel endpoint information on it.
Once the tunnel between the Hub Edge and the Super Gateway is restored, the spokes sites subsequently assigned to this Hub Edge will receive the Cluster info with the missing endpoint information, hence the tunnel creation will fail.
Impact / Risks
This issue can result in a Spoke Edge not connecting to any or all of the assigned Hub Edge Clusters with an impact on customer traffic routed for the Hub Edge Cluster.
Resolution
The problem has been addressed through internal ticket ID #74316.The fix for this issue is found on the following VMware Gateway releases:
On release 4.3.1, this fix is included with the 4.3.1 Gateway Roll Up #7: R431-20220509-GA which was released on 05-11-2022 and all later Gateway builds in this release train.On release 4.5.1, this fix is included with the 4.5.1 Gateway release R451-20220513-GA.On release 5.0.0.2, this fix is included with the 5.0.0.2 Gateway Roll Up #2: R5002-20220506-GA which was released on 05-12-2022.
Workaround
The customer can contact VMware SD-WAN Support and a Technical Support Engineer wil assist the customer in correcting the issue by rebalancing the Hub Cluster Members.
