HCX - Network outage observed in Layer two extended segments having NSX at source with HCX version 4.10.X
Article ID: 382811
Updated On:
Products
Issue/Introduction
The network outage is observed in regular L2E segments that have NSX at the source, where few of the VMs on the target site are not able to communicate with the source site, and the same VMs would be communicating with the VMs on the Target site without any issue.
Environment
HCX 4.10.0 & 4.10.1 with NSX-T registered at both source HCX and target HCX
Cause
Due to a change in NSX API calls in 4.10.x that the HCX manager uses, we have observed some changes in NSX Logical Switch(LS) properties that are causing the remoteRTR property to be removed from the Logical Switch level.
This occurs only when NSX is present in the source environment. Due to this, there are some configuration mismatches in the ports that inherit property from the parent Logical Switch and cause the packets to be dropped in the target HCX-NE.
Resolution
This is fixed in HCX 4.10.2. Refer VMware HCX 4.10.2 Release Notes
WORKAROUND:
Enable MON (Mobility Optimized Networking) for the stretched segment.
Additional Information
To help remediate the issue:
Enable MON (Mobility Optimized Networking) for the stretched segment. This would change the configuration to be applied on the port level, thus ensuring that the config is applied to each port individually.
Feedback
