my client is looking at Exchange 2007 High availability solution after the Ike hurrican took their entire exchange environment and it took 8 hours to fail over to the remote site. Client is following the CCR methodology. They require less than 1 hour turn around on the failover, and I am wondering about the following architecture: 1) WNS08 geocluster 2) use System Center DPM to help with synchronization and more efficient backup procedures. 3) CCR and SCR at remote site Questions: 1) Can this address the client SLA needs? 2) If we are making architecture recommendations, where is a good source that can talk about a end-to-end disaster recovery scenario using these technologies? 3) I found very little information about a Productionroduction hot backup datacenter scenario... what would be the difference in justscripting/automatingall the actions vs. a Hot backup datacenter? thank you for your help. Azin

Herere some documents which may help: Review File Share Witness Considerations section when you use GeoCluster <Forcing Quorum> Exchange Server 2007 with DPM CCR failover with SCR <Deployment>, flowchart of processes and decision making in this design SCR: Site Resilience with Standby Clustering <Scenario>

Would a Stretch CCR with WNS08 geo-dispersed cluster do automatic failover using the File Share Witness if the second datacenter can talk to the File Share Witness? My clients issue was the activation of the standby cluster SCR target. Since Houston and greater Houston area was completely without power, and without cell signal, the activation of the standby server was impossible until someone traveled out of the area to get a cell signal and send message to KC. Would the Geocluster environment in WNS08 support a automatic failover as is in the case of a single copy? thanks, Azin

If we put the FSW in a third datacenter, yes, it can do the automatic failover. You can find detailed info in Using a 3rd Site to achieve Automatic Failover section of here Notes: In this design, if third datacenter down, primary datacenter will continue to be active one, it can still communicate with the secondary [No error]. However, if we lost communication between primary/ secondary datacenter at this situation, the cluster service on both active/passive nodes will stop Notes: In this design, if the WAN is broken between primary/ secondary datacenter, primary datacenter will continue to be the active one as long as it can contact FSW; if theres a scenario that everything is just back online, both datacenters will try to access the FSW in the third datacenter, whichever do that first will become the active node and host all resources