Вы находитесь на странице: 1из 24

The installation: Starting from the very first node

So lets start the installation. If you remember, I mentioned that we have two steps to complete
the installation on both nodes. For now we will start doing the new instance installation in one of
the nodes. This node will be W2012SRV03.

1. With the SQL Server installation binaries available, click Setup:


2. The SQL Server Installation Center will open. Click Installation in the left menu, then
select New SQL Server Failover Cluster Installation:

3. A check will run in order to find possible constraints when installing the SQL Server
Setup support files. At the end of the check, click OK:

4. Now the installation will check for available updates, I recommend that you include those
updates into the installation. Click Next:

5. At this step, the setup support files will be extracted and installed, click Install:

6. Finally, we have all the setup files installed. Another check will run in order to validate if
problems might occur when SQL Server files is installed.

Best Practice:
Its recommended to have a clustered MS DTC resource, as well as a dedicated
MS DTC resource dedicated to each SQL Server group.
Here is a link to a good resource about this theme:
http://blogs.msdn.com/b/cindygross/archive/2009/02/22/how-to-configure-dtcfor-sql-server-in-a-windows-2008-cluster.aspx

7. At this step, just insert the product key and click Next:

8. Select I accept the license terms and click Next:

9. To install a SQL Server clustered instance, choose the first option SQL Server Feature
Installation. Click Next.

10. Here we have few options to install on Instance Features and Shared Features
sections.
For Instance Features pick the Database Engine Services item.
For Shared Features pick Management Studio (complete).Another important point
here is the directory that the shared tools will be installed. A good practice is have a
dedicated local disk to install SQL Server related files. In the image Im using the drive
C, the same that I have my OS. This is my lab, so I advise that this is not a good practice
for a productive system. Click Next.

11. At this step another check will run, this time to identify problems that might block the
setup, based on our choices of the last step. Just wait for its completion and click Next.

12. This is an important step: Here we will define the instance network name (one of our
prerequisites), whether we will use a default or name instance, the Instance ID and the
instance root directory.A few things to consider here:
On Detected SQL Server instances and Features on this computer section, we already
have one instance installed. So I had to use a named instance and change the Instance ID
to do not conflict with the existing one.
Based on the information above, to connect to our instance we will need to use:
SQL04\\DB.
Another important point is the Instance root directory: Its recommended to use a
dedicated local disk to install the SQL Server binaries. Avoid using the system drive,

a.k.a. drive C.Fill and verify all the points an click Next:

13. This step confirms if the disk space requirements are being met. Click Next:

14. Now this is related to the Cluster Resource Group name to be used. The installation
suggests a name, but you can change it.
This window also shows the reserved and already used Resource Group names. Change
the Resource Group name if you are not satisfied with the suggestion and click Next:

15. Now the disks! Another piece of the prerequisites has been shown. At this step, we will
have the information of all available storage to be used on our new clustered instance.
As mentioned before, I have only on available disk for this guide, so lets use this. Just
select the desirable disks and click Next:

16. Do you remember the requisite of an IP? Its time to use it! Just pick the network that you
will use, unmark the DHCP column and fill the address column with the value of the IP.
And Next:

17. Here is our last requisite: Service Accounts. Fill the information about the SQL Server
Engine and Agent service account (login and password).
Note: For a clustered instance, the Startup Type for the services should be as Manual.
The Cluster Service will manage this for us.
Click Next.

18. At the same step, we have another tab where we can define the collation to be used in our
database engine. By default it has Latin1_General_CI_AS.
For more information about collations, click here. This can be a very important choice! If
you have an special collation requirement, dont forget to select the right one during the

installation, otherwise you will have a hard work ahead to change this.

19. Now we need to choose the authentication mode of our instance.The options are either
Windows Authentication, which will take benefit from domain and local server logins,
or Mixed Mode, which accepts Domain/Windows logins as well as logins created and
managed by SQL Server.If you pick Mixed Mode a login called sa, member of the
Sysadmin role will be enabled. For this reason we need to specify the password for this
login.In the box bellow, we need to add all domain/Windows users that have access to the
instance and will be part of the Sysadmin fixed server role. You need to use the three
buttons to add/remove logins from this list:

20. Now its time to define the disk strategy.This step worth an entire article Im using
only one disk for demonstration purposes, but the recommendation is to use one isolated
disk for each one of the points.
The general rule here is: The most spread, the better!
We could use a layout like the following:

The only thing that we cannot specify directly is a place to store the non-clustered
indexes. This separation of the non-clustered and clustered index is a case aside.
Points to take attention here:

Isolate the TempDB in a fast disk. Remember: From SQL Server 2012 we can
store the TempDB in a local disk on clustered installations!
Place your data files in one disk and log files in another one.

Request the disks with an appropriate RAID level.

Pay attention to the partition offset and block size before the installation, even if
you are using a Windows 2008+ OS.

After consider all those points, set carefully the disks for each point and click on Next.

21. Here you have an option to send error reports or not. Click Next:

22. Another check will run to verify if the failover cluster installation will be blocked. After
the check, click Next:

23. Here we will be able to review all chosen options. Review and click Install:

24. The installation will begin wait for completion.

25. In the end you will have a confirmation about the success or not of each feature
installation.

You are done here! Click Close:

On this second part, we passed for all the steps to install the first node of a clustered instance. Of
course some points will be slightly different from your environment, but here are pretty much all
the steps to follow. After the completion of this installation you will need to add the other nodes
to this clustered installation and you will be able to see the SQL Server Role created into the
WFC Manager.

Вам также может понравиться