With last week's ship of vSphere 4.1, VMware made it clear that ESXi deployments are the 'best practice' platforms for vSphere 4.1, rather than continuing to use ESX servers. In short, making a plan for moving to ESXi is the current job of every Virtual Administrator working with VMware systems.
Data protection is a big part of any virtual deployment, and plans to migrate to ESXi must include your backup and recovery system. Vizioncore first introduced ESXi system support in vRanger Pro version 4.2. In this blog entry, we describe the key differences between how vRanger Pro operates when working with ESXi systems as compared to ESX systems.
Testing of vRanger Pro 4.5 with vSphere 4.1 has revealed no problems nor change in operations for the vRanger Pro product.
What is Different in vRanger Pro Support for ESX versus ESXi Systems?
vRanger supports ESXi using the vStorage API, which changes the way in which the backup process operates to move backup data from the source system to the repository. However, the backup engine and recovery process work exactly the same as with ESX systems.
In the illustration below, the difference in how vRanger Pro works with ESXi systems to collect data during the backup process is shown. Rather than injecting a binary in real-time as is done when working with an ESX system, with ESXi vRanger Pro makes a call to the vStorage API. This API call initiates the backup process by first triggering VSS and then creating a snapshot file to protect the VMDK file. The VMDK file must be available on shared storage to the vRanger Pro server, either on a SAN or iSCSI network connection, or locally on the network.
Today, Vizioncore does not support Active Block Mapping (ABM) on so-called LAN-free or shared storage deployments. ABM support for ESXi and shared storage deployments is planned for a future release. CBT is fully supported, however, as is the use of the Vizioncore VSS driver on ESXi Windows guests as a replacement for the VMware platform VSS implementation.
The backup image read from the VMDK file on an ESX system can be written directly over the network into an available disk repository, without passing through the vRanger Server. With ESXi systems, however, the backup image is read by the vRanger Server and is then written to an available storage location, which can be any local disk on the network as well as shared storage over SAN or iSCI. The difference is that the backup image is read and written by the vRanger Server itself, which therefore must have sufficient system capacity in terms of CPU and I/O to deliver the backup performance required.
How Does Restore Change with vRanger Pro Working with ESXi Systems?
The restore process for ESXi systems uses vStorage API, and is really no different than the restore process for ESX systems. vRanger Pro works with vStorage API to mount the VMDK and push the FLR or VMDK restore copy. A network-based restore using FTP is also fine to use. In short, there are no special restrictions or changes in the restore process.
Also, Recovery Manager for Exchange (RME) OLR for Exchange email objects working with vRanger Pro backup images is fully supported using ESXi systems, as well.