Reproduced with Permission from: http://telnetport25.wordpres.com

Like many Exchange Administrators whom have embarked upon a few recent Exchange 2007 migrations and deployments, I have chosen to use Windows 2008 as the host operating system and make use of Exchange 2007 CCR clustering capabilities in my overall system.

Although Windows 2008 and Exchange 2007 CCR clusters still make use of the MNS (Majority Node Set) cluster type – there has been significant changes to the underlying management and indeed architecture of the clustering service within Windows 2008.

Before I proceed if you are interested in (and I recommend having a look at) the following link: http://www.microsoft.com/downloads/details.aspx?FamilyID=75566F16-627D-4DD3-97CB-83909D3C722B&displaylang=en

Essentially it contains a number of word documents that give you a good overview of the changes and possible configurations of clustering within Windows 2008 – I especially recommend the following documents:

  • Windows Server 2008 Failover Clustering Architecture Overview.doc
  • Overview of Failover Clustering with Windows Server 2008.doc

Now that you have had a chance to look over the possible clustering scenarios (or indeed know you cookies already) I can get back to the article – Like many Exchange Architects before I deploy real life users (and to be honest at many stages throughout an Exchange deployment) I will run a Health Check of the entire environment using the Exchange 2007 Best Practices Analyser tool.

This is located in the Exchange 2007 Management Console – [ START -> Programs -> Microsoft Exchange Server -> Exchange Management Console ] and then navigating to (within the Console) [ Toolbox -> Exchange Best Practices Analyser ] – another quick tip to launch the Exchange BPA on a Windows 2008 Exchange 2007 server is to [START -> RUN ] and then type “EXBPA” and click on “OK

NOTE:

It is important to ensure that your copy of the Exchange BPA is using the most current XML definitions supplied by Microsoft. Microsoft updates them frequently with the latest data from their support groups – and indeed corrects “issues” with the BPA via this means.

Many of you will already know how to use the BPA and indeed how to execute a “Health Check” but for readers whom are not as familiar it is as follows:

Execute the Exchange BPA on one of the Exchange 2007 servers within your environment which contains the CCR cluster(s) (using the method above).

When the BPA has loaded (and you have moved past the BPA update screen) you will be asked to either “Select Options for a New Scan” or “Select a Best Practices scan to view” – choose the latter – see below:

BPAOdd-sc1

You will then be prompted for a Domain Controller to connect to – enter in the details of a suitable DC and then choose the “Connect to the Active Directory Server

BPAOdd-sc2

Permissions and network connectivity to your domain controller will be verified when successful you will then be prompted with the “Start a new Best Practices Scan” option – provide a name for your scan – then choose the Exchange servers that you would like evaluated (for the purposes of this article make sure you choose a Windows 2008 basedExchange CCR cluster) and ensure that you have selected “Health Check” option.

Verify the speed of your network and then choose the “Start Scanning Option

The BPA will then go away and verify your Infrastructure.

Now when the scan has finished you will be asked to view the report of the Best Practices Scan – confirm this option and you will be presented with a view that looks like the following:

BPAOdd-sc3

Now here you should address all issues which appear under the “Critical Issues” view – but for the purposes of this article we are interested in the Results for your CCR clusters which appear under the “All Issues” tab of the report – click on the “All Issues” entry.

Here you might find two issues which look like the following:

BPAOdd-sc4

When you expand the definitions of the above entries you will see the following:

Dedicated Heartbeat Priority:

BPAOdd-sc5

The full BPA definition of the above is available here: http://technet.microsoft.com/en-gb/library/aa997088.aspx

Quorum log too small:
BPAOdd-sc6

The full BPA definition of the above is available here: http://technet.microsoft.com/en-gb/library/aa995830.aspx

Now the point of this article is not the usage of the BPA tool itself – but the oddity that produces the above messages.

Upon further investigation of the above entries I have found that they either do not exist or indeed are set to the correct values as specified in the articles that have been suggested by the BPA.

I have searched many articles on the Internet and have found nothing that can explain as to why these values appear (under Windows 2008 clustering) as issues within the BPA (especially when they are either not needed, do not apply or exist already) so therefore the only conclusion that I could draw from this was that it is a “Feature” within the BPA itself.

As this was bugging me I contacted a friend in Microsoft whom confirmed to me that the above is indeed an oversight between versions of Windows 2003 and 2008 of with the current version of the BPA and that they intend to correct it very soon.

Therefore if you are running Exchange 2007 SP1 on a Windows 2008 CCR based cluster and are getting the above entries in the BPA Health Check report you can ignore them – but ensure that you update you definitions regularly to allow for the issues to be corrected.