4 key changes in Exchange 2013's high availability

Although Exchange 2013's underlying architecture is unchanged, admins will love automatic failover and other improvements

This week, I'm attending (and speaking at) my first Microsoft Exchange Conference (MEC), which was resurrected this year after being discontinued a decade ago and merged into Microsoft's TechEd conference. As an Exchange MVP, it's nice to rub shoulders with other Exchange folks, especially given the pending release of Exchange 2013 and its great features.

In my overview of Exchange 2013's changes, I mentioned there were no "large overhauls" in the area of high availability, although I did report the rewrite of the Managed Store. Scott Schnoll, a principal technology writer for Microsoft's Exchange Server team, reached out to say he disagreed with my comment. He also pointed out significant changes made to high availability in Exchange 2013, including updates related to database availability groups (DAGs). He suggested I take another look at it.

[ Get the details on what's new in Exchange 2013, what's new in SharePoint 2013, what's new in Windows Server 2012, and what's new in Lync Server 2013. | Stay abreast of key Microsoft technologies in our Technology: Microsoft newsletter. ]

I did -- but I stand by my original comment that there isn't a large overhaul when compared to the changes made from Exchange 2007 to Exchange 2010, where we went from several continuous replication options (LCR, CCR, and SCR) to the DAG continuous replication approach. In Exchange 2013, the DAG approach remains unchanged, as does the use of an active database with passive database copies that are maintained by log shipping and replay. There is no complete rebuild this time.

But Schnoll is correct that there are plenty of enhancements in Exchange 2013's high-availability capabilities worth your attention. After reading one of his recent posts and hearing him speak at MEC, I can better explain the value of those changes:

  • Multiple databases per volume: This lets you have both active and passive databases on the same volume, which is great when you consider the larger disks that Exchange 2013 can handle while still retaining low IOPS due to continued improvements to the information store and Extensible Storage Engine (ESE).
  • Auto-reseed: If a disk fails and has one or more database copies on it, Exchange can reseed to another spare disk (if available) through the replication service. The service can remap a spare and reseed. If there's more than one database to reseed, Exchange can do them in parallel.
  • Lagged copy enhancements: One change is the new Safety Net, which replaces the transport dumpster and provides a similar service of storing copies of messages successfully delivered to the active database of a mailbox server. (You can configure how long Safety Net holds these messages.) Safety Net is taking over the responsibilities of shadow redundancy in a DAG environment in that there is no need to keep another copy of a delivered message in a shadow queue while it waits to be replicated from the active to the passive database. That copy already exists in the Safety Net, so the shadow queue is redundant. One of the best features in Safety Net is the ability to activate a lagged copy: You can toss your transaction logs and mount the copy and let the Safety Net bring that database current.
  • Site resilience: This is by far my favorite new feature. Many Exchange admins have complained that, in data center failovers, you have to manually intervene and perform a site switchover. Not only does Exchange lack automatic failover between sites that experience a loss of quorum due to WAN outages and so forth, its manual switchover process is frustrating to work through. Exchange 2013 changes that: You can now have an automatic failover of your primary data center. There are some design conditions to take into consideration -- such as having three locations (two for Exchange servers and one for your witness server) -- but having the possibility of automatic failover is just awesome. If you do the manual switchover, Exchange 2013 reduces the effort to three simple steps.

Exchange 2013's high-availability enhancements also include storage recovery improvements and changes in best copy selection to ensure the failover only occurs to a system that is healthy.

This story, "4 key changes in Exchange 2013's high availability," was originally published at InfoWorld.com. Get the first word on what the important tech news really means with the InfoWorld Tech Watch blog. For the latest developments in business technology news, follow InfoWorld.com on Twitter.

Copyright © 2012 IDG Communications, Inc.

How to choose a low-code development platform