Exchange 2010 Backups? I don’t need any stinkin backups?!?
Exchange 2010 backups..do I really need them? This has started to become an interesting topic and one that Exchange and Storage administrators would have never thought of with previous Exchange versions. But, with the advent of DAG copies in Exchange 2010 it definately seems to be a hot topic within the Exchange community.
Microsoft tells us that going ‘backupless’ or what is known as ‘Exchange Native Data Protection’ is now becomes an option. With 3+ copies of the data, Microsoft says that backups become optional and adding an additional lagged copy with help provide the protection from logical corruption that is usually more likely to occur than physical corruption worries of the past. Lagged copies of the data are different from HA copies as they are designed never to be automatically activated, but are they solely to mitigate the risk of logical corruption. Microsoft explains this concept here: http://technet.microsoft.com/en-us/library/dd876874.aspx
Obviously, in a backupless Exchange 2010 environment we have to worry about log truncation, and that is where CRCL (Continuous Replication Circular Logging) comes into play. With CRCL on, logs get truncated and one could rely on their DAG copies and Lagged copies.
Quick VSS copies are still very much possible in Exchange 2010 and I honestly tend to see more of my customers staying with VSS methods of snap copies, especially if they are using great products such as EMC’s Replication Manager or other great VSS aware products from our other friends. These products have a long history of managing point in time copies and have become the rapid recovery tool of choice for customers.
In fact, we are just about to release a series of whitepapers that will showcase EMC’s solution using EMC Replication Manager for rapid VSS snap copies of Exchange databases. One solution will focus on a 2 copy DAG using EMC Replication Manager/TimeFinder on the EMC VMAX, and other will focus on a 2 copy DAG on the EMC CLARIION CX4-480, with EMC Replication Manager along with EMC Networker presenting the copy to a Data Domain 690 de-dup appliance. For the latter scenario, we called this a tiered approach where a requirement could be not only to take rapid VSS copies of the data for rapid recovery, but be able to keep the VSS copies for longer periods of time by being de-duplicated onto an appliance like the Data Domain. See below:
I think it comes down to the fact if your implementation has requirements for 3+ copies of the ExchangeI think it comes down to the fact if your implementation has requirements for 3+ copies of the Exchange data and if you want to deploy optional lagged copies from logical corruption that also need additional servers to support. Just remember that lagged copies can data and if you want to deploy optional lagged copies from logical corruption that also need additional servers to support. Just remember that lagged copies can only lag up to 14 days, so if your deleted item/single item retention and lagged copies are not kept beyond 14 days then you cannot recover past that point in time which can be a big deal for many customers that need to retain data for longer. Also, lagged copies can often require additional Exchange servers to add to the mix and it does mean you have to manage more copies, even though these are not copies that would be automatically activated.
I think overall, the backupless/Exchange Native Data Protection option certainly brings a new element to the Exchange 2010 architecture discussions and different options for providing recovery when bad things happen. VSS based options (hardware and software based) are still heavily used options to protect Exchange data that have been popular since the Exchange 2003 days and have increased with features and functionality into Exchange 2010.