DISASTER RECOVERY: SCR OR SCR WITH STANDBY CLUSTERING?
Posted on 2008-11-13
Microsoft recently helped me with a straight forward SCR solution for offsite replication of Exchange 2007, to be available and employed only in the event of a disaster in our office which disabled our production environment. SCR is designed so that there is always 50 log lag meaning that in the best of situations I would always lose 50 megabytes of the last transcactions. This is to ensure that corruption that might have occurred on the source database is not read into the target.
Microsoft has also published another solution that mixes SCR with standby clustering, that is more expensive in terms of Windows 2008 enterprise and Exchange 2007 licenses that I do not have as well as the time it will take to redo the current Exchange 2007 installation (on Windows 2008 standard, all servers on the same box, only two user accounts at the moment thankfully). However, this solution obviates the 50 log lag and apparently can be deployed faster in an emergency. I understand that mounting the mail databases in the straight-forward SCR scenario would take about an hour while in the mixes scenario abuout fifteen minutes.
The question is, if anybody has experience with these disaster recovery options, is the straight forward SCR solution adequate in an office where the 50 megabytes of lost transactional data would be spread over 165 users and the hour or so it took to get the target system operational was acceptable?