VMware SRM with remote host and remote array.

[seanhsmith]

Any issues with the following remote site scenario from a VMware vCenter or SRM perspective.

In this scenario. three sites are managed under 1 vCenter and recovered to the DR facility using SRM and vCenter.  the main question mark is around the role of the WAN/Latency and bandwidth in this situation.

vCenter 1  (At Primary site with remote  offices under management)

• Single vCenter
  
• SRM server
  
• Cluster 1
  
  • 5 Hosts
    
  • NetApp Storage replicated to DR facility synchronous
    
• Cluster 2 (Remote Site 1 with WAN connectivity)
  
  • 2 Hosts
    
  • NetApp Storage (at remote site) asynchronous replicated to DR facilit
    
• Cluster 3 (Remote Site 2 with WAN connectivity)
  
  • 2 Hosts
    
  • NetApp Storage (at remote site) asynchronous replicated to DR facility
    

vCenter 2 (DR facility)

• Single vCenter
  
• Single SRM
  
  • 5 Hosts
    
  • NetApp Storage Array recieving replication from primary site and  two instances of remote site storage.
    

Points of discussion.

1. Latency with hosts under vCenter issues - I am happy this is not an issue based on http://blogs.vmware.com/performance/2012/06/vcenter-50-performance-in-remote-offices-and-branch-offi...
   
2. Latency with SRA -> Array communications with the remotes. (Any thoughts on this requirements?)
   
3. Any other potential issues?
   

Any feedback appreciated?

[seanhsmith]

As additional information I found the following.

The SRM release notes have the following caveat.

http://www.vmware.com/support/srm/srm-releasenotes-5-1-0.html

Recovered VMFS volume fails to mount with error: Failed to recover datastore.

This error might occur due to a latency between vCenter, ESXi and SRM Server.

As I understand it this would not be a problem in the failover scenario because all the DR vCenter hosts etc are in the DR facility.  It could however be a problem in the failback scenario as the infrastructure is dispersed.

Any further comments?