In the previous post HCX & Virtual Machine Power State, we explored HCX handling of the various virtual machine power state scenarios.
For this post I am exploring various vSphere operations and their implications for HCX services.
The following terms are used:
Transfer phase: In this context it refers the Base VM sync (after setup/initializing, but before the switchover). The transfer phase also includes the Bulk/RAV’s “Waiting for Switchover” state (seen during scheduled migrations, this is when virtual machine changes/deltas are replicated at intervals.
Switchover phase: The orchestrated operation of bringing the vm live on the destination(for Bulk/DR) or transferring the memory state (for vMotion/RAV).
⚠️ If you haven’t already, please take a moment to read my disclaimer!
| HCX vMotion | HCX Bulk Migration | Replication Assisted vMotion (RAV) | OS-Assisted Migration (OSAM) | HCX Disaster Recovery | HCX Network Extension | Remarks | |
| Source vCenter Reboot | 🆗 During the transfer phase, but not recommended. ⚠️ HCX vMotion transitions immediately switchover. | 🆗 During the transfer phase. 🚫 Causes switchover to fail. | 🆗 During the transfer phase. 🚫 Causes switchover to fail. | 🆗 During the transfer phase. 🚫 Causes switchover to fail. | 🆗 During the transfer phase. N/A | 🆗 Does not affect NE forwarding. | ⚠️ vCenter Server should be online and fully functional for new HCX configurations. |
| Source vCenter Upgrade | ⚠️ Not recommended. 🆗 during transfer phase but HCX vMotion switchover cannot be scheduled. | 🆗 during transfer phase. ⚠️ will error on switchover, use schedule. | 🆗 during transfer phase. ⚠️ will error on switchover, use schedule. | 🆗 during transfer phase. ⚠️ will error on switchover, use schedule. | 🆗 during transfer phase. Recovery does not use source vCenter. | 🆗 Does not affect NE forwarding. | HCX ideally continues to function without hiccup with standard vCenter upgrades where the UUID and database contents don’t change. |
| Intra-Cluster vMotion | 🚫 Not Available | 🆗 During the transfer phase. The replication is resumes on another host after a brief pause. | 🆗 During the transfer phase. The replication is resumes on another host after a brief pause. | N/A | N/A | N/A | ⚠️ vMotion of the HCX Service Mesh appliances is supported but discouraged. ⚠️ A vMotion of the IX always displays “Virtual ethernet card ‘Network Adapter 1’ is not supported.” but it does not disrupt the IX or vMotion operation. |
| Storage vMotion during VM Migration | 🚫 Not Available | 🆗 During the transfer phase. The replication resumes on another host after a brief pause. | 🆗 During the transfer phase. The replication resumes on another host after a brief pause. | N/A | N/A | N/A | ⚠️ Storage vMotion of the HCX Service Mesh appliances is supported but discouraged. |
| HCX vMotion | HCX Bulk Migration | Replication Assisted vMotion (RAV) | OS-Assisted Migration (OSAM) | HCX Disaster Recovery | HCX Network Extension | Remarks | |
| Maintenance Mode (ESXi) | The host will not enter maintenance mode until all the VMs are migrated or powered off. An existing vMotion needs to complete. | 🆗 The replication is resumes on another host after a brief pause. | 🆗 During transfer the replication is re-configured on another host. | N/A | 🆗 During transfer the replication is re-configured on another host. | If the HCX-NE is present on the host it will be migrated. | |
| Ungraceful Reboot (e.g. ESXi loss of power) | 🚫 vMotions will fail. | ⚠️ Transfers will be halted. Switchovers will error. | ⚠️ Transfers will be halted. Switchovers will error. | N/A | ⚠️ Transfers will be halted | Forwarding will be disrupted until NE is started. | |
| Workload VM Disk addition /removal | 🚫 Cannot edit VM settings once relocate starts. | 🆗 The replication cycle will detect disk changes. | 🆗 Replication detects CPU/MEM changes during transfer. 🚫 Cannot VM edit settings once relocate starts. | ⚠️ Changes may not take effect and may lead to migration failure. | 🆗 Replication detects CPU/MEM changes during all replication stages. | N/A | |
| Workload VM Disk Increase | 🚫 Cannot edit VM settings once relocate starts. | 🚫 Cannot change disk size for disks currently being replicated. | 🚫 Cannot change disk size for disks currently being replicated. | ⚠️ Changes may not take effect and may lead to migration failure. | 🆗 Solution: Recover the VM to an isolated network, Power-off the recovered VM. Resize original at src. Resize recovered VM at dst.. Reconfigure the Protection with the powered-off VM as the seed. | ||
| HCX vMotion | HCX Bulk Migration | Replication Assisted vMotion (RAV) | OS-Assisted Migration (OSAM) | HCX Disaster Recovery | HCX Network Extension | ||
| Workload VM CPU MEM Changes | 🚫 Cannot edit VM settings once relocate starts. | 🆗 Replication detects CPU/MEM changes during all migration stages. | 🆗 Replication detects CPU/MEM changes during transfer. 🚫 Cannot edit VM settings once relocate starts. | ⚠️ Changes may not take effect and may lead to migration failure. | 🆗 Replication detects CPU/MEM changes during all migration stages. | N/A | |
| Workload VM .vmx changes | 🚫 Cannot edit VM settings once relocate starts. | 🆗 Replication detects CPU/MEM changes during all migration stages. | 🆗 Replication detects CPU/MEM changes during transfer. 🚫 Cannot edit VM settings once relocate starts. | ⚠️ Changes may not take effect and may lead to migration failure. | 🆗 Replication detects CPU/MEM changes during all migration stages. | N/A | |
HCX ⦿ DESIGN 
[…] CPU). Bulk migration can tolerate some amount of some refactoring of the source virtual machine (tell me more). This makes HCX Bulk migration a good option for any scenarios some planned application downtime […]
LikeLike