Figure 2

You have now seized all 5 of the FSMO roles for the lab PDC, and you have cleaned out all evidence of the production DCs from our lab DC. Now we need to remove the lab DC object from our production DCs. Although we have 2 production DCs it is only necessary to remove the phantom DC from one of them. I suggest connecting to your PDC to do this. So log onto the production PDC as the domain administrator and open a command prompt. Follow the steps above for removing DCs using ntdsutil, but this time remove the lab domain controller. Do not seize any FSMO roles since the computer you are logged onto already has them all. Then, open the DNS snap-in. In the left pane navigate to the forward lookup zone for your domain name. In the right pane double-click one of the Name Server type records. You will see a dialog box opened to the Name Servers tab. You will see your labdc server as well as both the production DNS servers listed there. Delete the lab DNS server.

At this point you need to modify your DNS setup on the lab domain controller. Your DNS server is going to show all the zone information that you have in production regardless of when you installed DNS because that zone is stored in Active Directory, and has been replicated over from production. Just go into the DNS management snap-in and delete all of the forward lookup zones as well as the reverse lookup zones. When you are done, add the forward and reverse lookup zones you need for your environment making sure to select Active Directory-integrated zones. When the new zones are created they will contain only the SOA, A, and NS records for your new lab domain controller. All the computers we add to the lab domain from here on out should be able to add themselves automatically.

We are pretty well on our way at this point. We have cloned our production Active Directory database, and have successfully removed all phantom domain controller information from our 2 domains. So production is back to functioning normally. All we have to worry about at this point is the lab, but we already have the backbone for creating this entire network. Next we will bring online our other domain controller.

I am not going to run through every detail of bringing up this next domain controller. The important thing is to just set it up with the roles we want it to have. So first install Windows 2000 Server. Then join it to the domain.com domain, assign it a static IP address, and run dcpromo. Once it is a domain controller we will want to install DNS. After we have that set up we can go ahead and install SUS, and you can use my Windows Patch Management article for the specifics involved in that installation. The last role we need this box to fill is that of our central Norton Antivirus server. You should consult the installation guide for the specifics of installing NAV Corporate Edition.

Bringing Exchange Online

Our 2 domain controllers are up and running in the lab. We are offering all of our important network services, except for email of course. I guess that gives away what our next step is going to be then- install our Exchange server. First we start by installing Windows 2000 Server, running all updates, and setting a static IP address. The next step is to install Exchange Server 2000. Once again, this is not a "How To Install Exchange" guide, but I will be covering the points that are important to our specific situation. For a general guide to installing Exchange 2000 you can try this, this, or this.

We bought a large IDE drive for this server so make sure that is installed and formatted before installing your copy of Exchange Server 2000 from your MSDN subscription. Also, use ntbackup to create a backup of both your public and private information stores on your production Exchange server. You will need this later to populate the lab Exchange server with all of your mailboxes and public folders. There is no need to perform any of the domain prep or forest prep actions described in the last link above since we have replicated Active Directory from a domain where these actions have already been performed. However, when you attempt to install Exchange normally you will receive an error stating that the server object already exists in Active Directory (I think we all know why that is). So according to Microsoft we can run the installation using the /disasterrecovery switch to get the server back online, and then use the ntbackup of our production server to get our lab server completed as a mirror image of production.

When you get to the "components" page of the installation routine make sure you've selected Microsoft Exchange 2000, the Messaging and Collaboration component, and the Management Tools. Then change the installation path to the big partition you made on the new IDE drive, and then proceed. The rest of the process should take a bit of time, but not much effort. Grab a drink and then figure out how to move that 25GB backup file of your production information store over to the lab Exchange Server. In my case I used one of those Maxtor OneTouch Backup drives we had around for emergencies (no I don't own stock in Maxtor). It's just a big IDE drive in an external enclosure with a USB interface. I copied the NTBackup file to that and then plugged that drive into our lab Exchange Server.

So, we've completed the Exchange installation and we have our backup ready to be restored. Before you do this you need to apply service packs to get your lab Exchange Server to the same patch level as your production one. Otherwise you will not be able to mount the Information Store after you restore it. So in our case we download Exchange Server 2000 Service Pack 3, and run the update.

Now, you're ready to restore. In order to perform the restore make sure that your Microsoft Information Store service is running, and that your public and private stores are dismounted. Open the Exchange Manager --> expand First Storage Group --> Servers --> ServerName --> OrganizationName. Right-click the Mailbox store and select Properties. On the Database tab check the box that says "Allow this database to be overwritten by a restore." (see Figure 3) Do the same for the Public Folder store.

Figure 3

Now open NTBackup, and start the restore wizard. You will need to Import the backup file you have from your production system so that NTBackup will allow you to restore from it. So on the first page of the restore wizard select Import File (see Figure 4), and then select your backup file. When NTBackup is done importing it the file will appear in the window below the heading "What to restore." Select it and click Next. The next page should automatically populate the location you are going to restore to with the name of your email server, but input this information if it does not (see Figure 5). Select a place for temporary files to be stored, and check the Last Backup Set check box since we have no more files to restore after this one. I usually choose not to automatically mount the database after the restore, but that's personal preference. Feel free to check that box if you like.

Figure 4

Figure 5

The restore should take anywhere from a few minutes to a few hours depending on how big your information store is, and how fast your computer is (your disks will be the main bottleneck). For my ~25GB store the restore job took just under 7 hours to complete. After the restore you are going to have to actually get your hands dirty a bit. Your databases are not going to be in a consistent state, and are therefore not going to mount properly. Follow the instructions below to bring both databases into a consistent state. This total process may take a full day or more. Step number 3 in my environment took about 4 hours to complete. The public folder store was smaller and took less time, but then defragmentation takes a while as well. So it's a good idea to plan on doing something else while these tasks complete.