I got an email with this question – what happens when you create a DAG in 2010 SP1? The person who emailed was under the impression that a CNO gets created as soon as a DAG is created in EMC. Does it? Let me break down the process.

Once you create a DAG in the EMC, nothing much happens, atleast that is visible to an admin (unless you start looking with ADSIEdit). To explain what happens, I have created a DAG named DAG1 in my lab.

A CNO is NOT created in AD, once a DAG is created in EMC. Some admins expect a CNO and thinks that something has gone wrong. It will only be created once you add the first mailbox server.

The FSW that you have specified doesn’t get created now as well, only when the servers are added.

I added EXCH1 to my DAG in the lab. Now, things start to happen! A CNO with the DAG name gets created in the Computers OU in AD.

Windows Failover Clustering gets installed on the mailbox server. A cluster with the DAG name gets created and the quorum model is set to majority node set.

Once the second mailbox server (EXCH2 in my case) is added, windows failover cluster gets installed. The the cluster quorum mode changes to node and file share majority model. This switch happens automatically as servers are added and removed from a DAG.

The FSW has some contents in it now.

This will be removed if the number of DAG members becomes an odd number. Hence, as soon as you remove one of the server or add a new one, share folder will be empty. Nothing to panic, that is by design. The changes to the quorum model and FSW is controlled by Exchange and no changes should be made by an admin in normal cases.

All actions related to the DAG are logged in ExchangeSetupLogsDagTasks.

Hope the article shed some light as to what exactly happens when a DAG is created.

Until recently, the maximum round trip latency you should have was 250ms for a DAG across datacentres, to have it supported by Microsoft. This has now been doubled and the new supported latency is 500ms. This was stressed in the Technet event held in Australia recently.

Microsoft still recommends to have a lower latency to have a better performance. The slide below from the Technet video explains the new support stance.

I had written before about Kemp technologies load balancer online sizing guide, read here. Kemp has released a 30 day evaluation version of their virtual load balancer named LoadMaster.

It works with VMware ESX, ESXi, VSphere, Workstation, Server, Player and Microsoft Hyper-V. Kemp provides the documentation for setting up the load balancer as well. There is no excuse not to trial it

Download the evaluation version here

I had written before about how Exchange Shell command log works in 2010 RTM, read here. Things have changed in SP1. The feature is enabled by default and logs all commands that have been executed in the session.

In order to access it, launch EMC, navigate to View –> View Exchange Management Shell Command Log.

Clicking on it opens the Command Log window. All commands executed are shown in the log window.

You can stop logging, modify the number of logs to store, clear log, copy commands and export the list of commands.

Taking a close look at the commands is a good way of learning to script.

Microsoft has released Update Rollup 5 for Exchange 2007 SP3. As usual, it is a cumulative update and can be applied without installing previous rollups. It should only be applied if you are running atleast Exchange 2007 SP3.

Forefront should be disabled using fscutility /disable before the setup and enabled after using fscutility /enable.

Description of UR5 here

Download the rollup here

With the release of cheaper load balancers (which are Microsoft certified as well), most of the companies tend to use a hardware load balancer for load balancing Exchange 2010 traffic. There are three options that work for routing the traffic. They are:

Source NAT (SNAT) : The packet from the client hits the load balancer, which passes the traffic onto one of the CAS servers after changing the source IP of the packet (from the client’s IP to the LB IP). CAS responds back to the load balancer, which sends the packet back to the client after changing the source IP.

This is the Microsoft recommended option for routing. All traffic (in & out) flows through the load balancer.

Direct Server Return (DSR) : The packet from the client hits the load balancer, which passes the traffic onto one of the CAS servers, without making any change to the source IP and the CAS server responds directly to the client. This means that the CAS server needs to have the load balancer’s IP configured as an additional one.

Load Balancer Default Gateway (LBDG) : In this method, the CAS servers have the LB IP as the default gateway.

The below screenshot (taken from TechEd video) explains the three options.

Finding the update rollup version you have installed on your Exchange 2010 boxes seems like a simple task. It is, if you know where to look for! There are a number of questions posted in Exchange forums where admins get confused with what they get from console/shell. The common places were an admin will look for displays only the major version numbers, like 2010 RTM, SP1 etc, and not the update rollups.

First place to look to find the build number for various update rollups is this Technet Wiki page. If you are running the latest rollup, the build number should be 14.1.339.1. Now, how do you find this number in your Exchange box?

If you have only a couple of Exchange 2010 servers, easiest is to go to Control Panel –> Programs and Features –> View Installed Updates.

Another option is to navigate to the Bin directory and select the properties of “Exsetup”.

The File version in the “Details” tab gives you the build number.

Yet another option is to open up Exchange Console and click on “Help” menu –> About Exchange Server 2010.

This brings a small window, which gives you the build number.

Run (Get-Command Exsetup).FileVersionInfo from Exchange Shell to find the build number easily.

This is good, if you only have a couple of servers. But, what if you are in a large environment and want to get this info? There are a number of PowerShell scripts which does the job. Check this script as an example.

I was asked this question today. Why did Microsoft pick “16” as the maximum number of servers that you can have in a DAG? Having 16 servers in a DAG means that you can have a maximum of 16 copies of a database (one active and 15 passive). Isn’t it a bit too much to have 15 passive copies of a database? How can you justify this figure?

These were the questions that bugged an exchange admin that I met. To make it clear, the Exchange team didn’t go for a “magic” number for the maximum number of servers in a DAG. The limit is imposed on the DAG by the underlying Windows failover clustering, on which DAG is based.

The Windows failover cluster can only support a maximum of 16 nodes. This limitation kind of flows into DAG. So, though a DAG can only have a maximum of 16 servers, the limit actually comes from the Windows failover cluster.

There you go, the story behind the magic number

If you haven’t been aware, it is time to upgrade your Exchange 2010 RTM boxes to SP1. The support for 2010 RTM is ending on Oct 11, 2011. This gives you three weeks to plan and upgrade all your Exchange boxes, before you go out of MS support.

More info @ source

I tried installing AD on Windows Server 8 DP build, explained here. Everything went fine as long as I went with the default values for the forest & domain functional level, which is Windows 2008 R2.

Just being curious about any new functionalities that a pure Windows Server 8 functional level might expose, I changed the forest & domain functional level to Windows Server 8 during dcpromo. To my surprise, it didn’t like it! The wizard threw an error when the prerequisite check was done. The error is shown below.

DCPromo Failed With Code: 77. The Specified value ‘5’ Is Not Valid For The Argument ‘DomainLevel’

Is this a bug Microsoft? Or is it looking for some prerequisites that needs to be installed to have a pure Windows Server 8 domain? If so, why doesn’t it say what is required?

As soon as I changed the functional levels to 2008 R2, installation went smooth. Anyone else having the same issue?