Database Reference
In-Depth Information
Chapter 4. Virtualizing SQL Server: Doing IT Right
This chapter focuses on the things you need to know and do as you start down the path of
database virtualization. The advice given in this chapter takes a very conservative
approach, with the end goal of helping you avoid the common traps and pitfalls
encountered when you first virtualize a production SQL Server database. Topics
covered include the following:
Documentation
The implementation plan
The importance of obtaining a baseline
Additional considerations
A bird's-eye view of the implementation process
Doing IT Right
Our experience has taught us that the best place to start down the path of database
virtualization is to read the documentation. The first thing many DBAs do when a new
version of the database comes out is to install it and start using it. (In a nonproduction
environment, of course—no DBA is going to deploy a new version of the database,
including a database patch, without first testing it.)
The problem is that those same DBAs don't always circle back and do a complete read
of the documentation from front to back. This is further compounded by the fact that
vSphere is easy to install and use right out of the box—it lulls you into thinking you do
not need to read the documentation. A strong word of caution is in need here: What has
worked up until now in your virtualization infrastructure will not necessarily work
when you put the demands of a production database onto that environment!
Tip
Read all the documentation from all the vendors. That includes VMware,
Microsoft, the network vendor, and especially the storage array vendor—in
particular, their SQL Server Best Practice Guides.
A virtualized database is sitting on a shared environment, even if it is only shared with
other SQL database systems. Therefore, it is very important that you take the time to
read all the different vendors' documentation. We place special emphasis in this topic
on reading the storage array documentation. Our experience has taught us that over 80%
of the problems with virtualization implementations occur at the storage layer. Here are
some examples:
