Failed to attach vif uuid to network nsx.LogicalSwitch:xxxxx in an SRM/NSX-T environment

Products

• VMware NSX

Issue/Introduction

In the environment, SRM and NSX-T are used.

• There are two NSX installations, one per site: Source and Destination.
• When performing SRM test or failover, they fail and the step 4 (Snapshot creations) of the SRM reports contains error like:

An error occurred during host configuration. An error occurred during host configuration. An error occurred during host configuration.

• SRM network mapping is done correctly.
• On the destination vCenter, some VMs are not attached to the expected destination NSX-T segment.

• In the NSX-T Manager (Destination) logs, there are errors reporting the VM cannot be attached to these segments as they don't exisits:

Failed to attach vif uuid to network nsx.LogicalSwitch:xxxxx *Where xxxxx will be the id of an existing segment on the source environment.

• When the issue occurs, you will be unable to attach any NSX-T segments to VMs (on the destination ESXi hosts), until the NSX-T services (nsx-opsagent and nsx-proxy) have been restarted on these destination hosts.

Environment

VMware NSX

Cause

There is an enhancement needed in the NSX and SRM interop.
Also an important amount of logical port creation on an existing NSX Segment alter the stability of the NSX services (nsx-opsagent / nsx-proxy) running the ESXi hosts.

Resolution

Currently there is no resolution.

Workaround:
By reducing the number of VMs migrated by SRM into smaller group, the migration can work. As example:

1. Create X groups of 50 or 100 VMs.
2. Start the migration of the first group .
3. On the destination, on all ESXi hosts, restart NSX-T manager services:  nsx-opsagent and nsx-proxy services logged in as root:

/etc/init.d/nsx-opsagent restart
/etc/init.d/nsx-proxy restart

       4. Repeat steps 2 and 3 with the next group.

Additional Information

Impact/Risks:
The SRM migration will fail.