<img height="1" width="1" src="https://www.facebook.com/tr?id=1529264867168163&amp;ev=PageView &amp;noscript=1">

The ESE Consultants’ Corner: Fourth Edition

The Consultant’ Corner

Hello! My name is Mahmoud Magdy and I am honored to be hosting another edition of the ESE Consultants’ Corner. For this edition’s post, I chose to answer several questions I recently received that will benefit our readers the most.

Q: I have setup my CAS servers in a CAS Array but load balancing is not taking place. All of my CAS servers are included in the Array, what am I doing wrong?

As a good starting point, you should first read the excellent article that Exchange MVP Henrik Walther wrote on this topic at:


Listed below are some additional tips and tricks that will help you configure your CAS Array correctly.

  • Setting up a CAS arrays does not provide NLB services by itself, but rather it groups your CAS servers together. By default it will include all of your CAS servers in the local AD Site. You can only have one CAS Array per AD Site.
  • A CAS array is identified by its FQDN. An example would be casarray.domain.com
  • In order to provide Network Load Balancing services for your CAS Array, you will need to use an external NLB technique to load balance the traffic designated to CAS Array hostname. You can use Windows Network Load Balance or an external Hardware Load Balance solution.
  • After you have created your CAS Array, do not forget to change the RPCCLIENTACCESSARRAY property on all of the databases to point to the newly created CAS Array. By default the value will still point to the first CAS server created.

Q: I am trying to setup a disaster recovery site for my Exchange servers. I have 2 Active Directory sites, one located at my head quarters the other at the DR site. I only want my databases to be activated at the DR site if all of the mailbox servers in my main location go down. I have configured the Activation preference option but it is not working. What am I doing wrong?

The first step is to confirm that you are using Exchange 2010 SP1.

Database activation within DAG is managed by the Active Manager. Activation can either configured manually or you can allow Exchange to manage it.

Prior to SP1, Active Manager had its own algorithm to determine which databases should be activated. Several conditions were taken into account including, log replication health, content indexing health and replay queue length. There are several other conditions which can be found by reading the Exchange 2010 RTM documentation. The Activation preference is only used when the aforementioned factors do not apply.

With the introduction of Exchange 2010 SP1, Microsoft has made minor changes to how Active Manager determines when to initiate a database failover. If you have your databases configured with the lossless high availability option, the Active Manager will now use the Activation Preference as the primary factor to determine when to failover the databases.

Q: I have SP1 installed and the Activation Preference is set, but the automatic failover does not occur as expected. If I move them manually it works. How do you explain this behavior?

The most likely cause for this behavior is that all of your Databases are configured with the good availability mount dial setting. This is the default setting when you create a new database. In order to have the Activation Preference take priority, you must set the databases to use the lossless mount dial setting.

When you move the databases manually with a cmdlet it will use lossless by default and this is why it works using the manual method.

Q: I am planning on creating archive mailboxes for my users and am considering using 15k rpm SCSI hard disks?

Before I begin, I would like to add that anything I write in this column is my humble consulting opinion.

Archiving is meant to host historical emails for a long period of time on cheap storage where it can be retrieved later. Since Exchange 2010 SP1 now provides you the ability to create the archive mailbox on a different database, you can now store this data on cheaper disks.

The above statement may not be valid for organizations that have real legal or compliance requirements because in those situations keeping all sent and received emails will require higher performing drives.

Thanks for reading my post and until next time.