Le Gia Vinh’s Blog

This is one of excellent papers about Virtualization. The full paper is from here.

Source

VMware’s new “Site Recovery Manager”. What’s the best way to describe this? See it as HA for two different locations - a Primary and DR location. For it to work you need some kind of replication (synch or asynch the choice is yours). The replication of the VM is not being done by VMware but by your SAN, iSCSI or NAS vendors replication technology. SiteA can protect SiteB, and SiteB can protect SiteA. So this does offer up the opportunity to share DR costs - like two universities who can recipocate each others resources for DR purposes. VM’s can continue to have a “home” location where they normally reside.

The Vi-Client has a plug-in extension that speaks to the SRM Service - and orchestrates the whole process. Creating “Protection Groups” (groups of VMs you need to have DR for) which are very closely linked to “LUN Groups” (where their storage is located). The SRM plug-in handles the “mapping” of Virtualcenter object in the VC environment in SiteA to SiteB such as folder and reasource location. Additionally, their settings that handle the virtual switch port group, and scripts to allow for changing of IP settings when VM stops running at SiteA, and needs to run at SiteB with different IP configurations. Essentailly, the SRM service at the primary location handles the configuration, and then pushes its out to backup location. The feeling is of one VC when in fact two exist which are synch’d together by the SRM software. The VMs created at the second location are referred to as “shadow VMs”, which exist in a kind of “shadow” environment at the DR location.

Lastly, you can test your DR running different “recovery plans” which allow you to workout that given failure X, the plan Y would be the best approach. The product was actually announced yesterday.

Link:

https://www.vmware.com/company/news/releases/srm.html