By leveraging VSP One Block High End, enterprises can unlock greater efficiency, reduce infrastructure complexity, and future-proof their virtualization strategy.

Accelerate virtual machine migration to OpenShift with storage offload

Hitachi has partnered with Red Hat to introduce a powerful storage offload feature in the Migration Toolkit for Virtualization (MTV) Operator, available starting with MTV version 2.9 as Tech Preview. If you are planning to migrate VM workloads from a VMware vSphere cluster to OpenShift Virtualization, and both environments are backed by the same VSP One storage system, you can take advantage of this feature to dramatically streamline the migration process. 

Learn more about this collaboration and feature from the following article: https://www.hitachivantara.com/en-us/blog/replatform-faster-openshift-vsp-one-storageoffload 

The following lists some of the main benefits of this storage offload migration: 

• Up to 10 × faster migration: Internal testing shows up to 90% reduction in migration time – reducing a 10-hour process into a 1-hour operation.
• No IP network dependency: VM volume data never leaves the VSP storage, freeing up network bandwidth and reducing latency.
• Minimizes compute host resource usage: Less time spent on migration tasks, preserving CPU and memory for critical workloads.

There are several options for migrating VMs from VMware reside on third-party storage or older generation storage to OpenShift Virtualization backed by VSP One Block High End enterprise storage. In this article let’s discuss following two options. For more migration options in detail, see this document:

https://docs.hitachivantara.com/v/u/en-us/application-optimized-solutions/mk-sl-410

Cross-storage, Standard VM Migration with Network

If you are migrating VMware VMs from third-party storage or previous-generation storage to new VSP One Block High End storage, the default standard method is to copy the VM data over the network as shown:

Cross-storage, Standard VM Migration with Network:

• This is the default MTV setting. Simpler to set up and configure
• MTV uses network data transfer to copy VM disks from the source VMware cluster to the target OpenShift Virtualization cluster.
• All disk data flows through the migration network, consuming bandwidth and CPU resources.
• Supports both Cold migration and Warm migration.
• Slow migration speed. For example: one VM with 1 TB can take about 3 hours over a 10 GB Ethernet link.

Cross-storage, VM Migration with Storage Offload

If you are migrating VMware VMs from third-party storage or older storage to the new VSP One Block High End storage, this is a significantly faster method.

The following is a list of the storage offload steps, after the MTV VM migration plan is executed:

1. An empty persistent volume (PV) is provisioned from VSP storage and temporarily attached to the source ESXi host.
2. MTV instructs the ESXi host to copy the VM data to the PV using XCOPY commands. However, because the source storage and destination storage are different, the XCOPY fails. When XCOPY fails, it falls back to the regular data copy within the ESXi host, using the Fibre Channel or iSCSI data path. This VM data copy is still much faster compared to the standard network copy.
3. After the VM data copy is complete, the PV is detached from the ESXi host.
4. The same PV is now attached to an OCP worker node.
5. MTV performs an image conversion to complete the VM migration.

As of MTV 2.10, only Cold migration is supported for Option 4. Warm migration will be available soon

Summary

Make sure to see this document for more detail:

https://docs.hitachivantara.com/v/u/en-us/application-optimized-solutions/mk-sl-410