Site Resiliency for Exchange 2007
This topic has come up twice in the same week. Quite a hot topic. We’ve talked about CCR and SCR in the past this topic will focus on options for site resiliency. This is where we have a primary and secondary center. The first item to understand is what type of recovery datacenter configuration do we have: cold, warm, or Hot. Also, what is the resource allocation at the recovery datacenter – are they dedicated or not dedicated to the site resilience solution. So what are the options for our datacenter.
Old Method of Site Recovery and Data Replication
Ship backups and restore
New Methods for Site Recovery and Data Replication
- CCR+SCR and /recoverCMS
- SCC+SCR and /recoverCMS
- CCR stretched across datacenters
- SCR and database portability
- SCR and /m:recoverserver
- SCC stretched across datacenters with synchronous replication
- 3rd party (replicate and boot from SAN, non-MS clustering, etc.)
In this scenario we have a warm, dedicated CCR+SCR. Warm because it requires manual steps to bring the secondary site up. Dedicated because the servers are dedicated to the services in each site.
A variation on this is a Hot, non-dedicated CCR+SCR with /recoverCMS within a stretched AD Site. Here the site is stretched and we have CAS/HUB and GC that can service both sites. All routing can flow to/from either site and you have a scenario now where both sites are providing all traffic. This may not be the scenario you want if you are looking at a primary site performing most of the traffic flow.
This scenario is for a hot, non-dedicated stretched CCR with a stretched AD site. The site needs to be stretched in order to support the cluster. The HUB/CAS/DC are across both datacenters in this scenario so a HUB/CAS/DC in one datacenter can handle the load of both centers. There are many details to each of these scenarios one in this one is that the File Share Witness (FSW) for the CCR cluster is on the Hub Transport in that datacenter. If it’s a full datacenter loss then both CCR clusters can’t recover since the FSW isn’t available and either is the Active Node. 
The final one I’ll cover today is the hot, non-dedicated CCR+SCR with/recover CMS across AD sites. In this scenario you may have noticed that the CAS3 and CAS4 boxes have been added to the scenario. These are only needed if you have CAS servers that don’t support SAN Certificates. If SAN certs are used you don’t need to drop in a server with the Cert into the new datacenter.
References:
SAN Certs - https://technet.microsoft.com/en-us/library/bb851505.aspx
https://www.exchangeninjas.com/CasCertMethod1
https://msexchangeteam.com/archive/2007/04/30/438249.aspx
Site Resiliency - https://technet.microsoft.com/en-us/library/bb201662.aspx
Webcast on Site Resiliency - https://msevents.microsoft.com/CUI/WebCastEventDetails.aspx?culture=en-US&EventID=1032381323&CountryCode=US