Red Hat flags Linux connectivity loss after VMware VM migrations with MTV

Red Hat has published a verified support solution for a failure mode that can leave a Linux virtual machine without network connectivity after it is migrated from VMware with Migration Toolkit for Virtualization (MTV). The solution headline says the guest’s network interface is renamed with an _bkp suffix and connectivity is lost.

That makes this more than a cosmetic naming issue. A VM can complete the transfer and still fail the first operational test that matters at cutover: reaching the network. Red Hat’s publicly indexed summary identifies the symptom, but does not expose the full cause or remediation steps. Teams should not assume that a completed migration means the guest is ready for service.

Treat network validation as part of cutover

For VMware-to-OpenShift Virtualization moves using MTV 2.11, we recommend testing a representative Linux guest before a larger wave and recording its interface names and expected network behavior before migration. After the migrated VM starts, verify the interface name, address, routes, DNS resolution and application reachability before retiring or disconnecting the source workload.

If the interface appears with _bkp, preserve the failed guest state and migration evidence rather than immediately making broad changes across the migration plan. The exact supported correction should come from Red Hat’s full solution or a support case, because the public summary does not establish which guest configurations are affected or whether one recovery procedure applies to every environment.

MTV exposes the evidence operators need

Red Hat’s MTV 2.11 migration guide says operators can view migration logs while a migration is running or after it completes. From the migration plan, the Virtual machines tab exposes each VM’s status and expandable detail, giving teams a place to capture the failed VM’s timeline before retrying.

The same guide says a migration in progress can be canceled for some or all VMs. A canceled migration plan can be restarted. That control is useful when the symptom appears in an early VM: pause the wave, inspect the migrated guest and logs, and avoid carrying an unverified network state into the rest of the batch.

The practical decision is straightforward: add guest-level network validation to the MTV cutover gate, retain access to the source until the migrated VM passes it, and use Red Hat’s verified solution as the support reference if the _bkp rename appears.