Storing Microsoft SQL Server Databases on the NAS 8000

would involve the use of the SQL Server Enterprise Manager in order to change the path to the database
data/log files. Failure to correctly update all occurrences of the IP address could result in a database
failure, or possible data loss. Whereas, if the network name were used, only the address in the Name
Server(s) would need to be updated.
mapped network drives
Using a "mapped network drive" is another method to access the NAS 8000, although drive mappings are
session dependent rather than session independent as with direct UNC naming. Mapping a network drive
assigns a drive letter to a NAS 8000 share. The 'connection' between the Windows server user and the
NAS 8000 is established (along with any security) when the drive is first mapped and whenever the
Windows server user logs onto the Windows server, if the mapping is "permanent". When mapping a
network drive, use the UNC name for the NAS 8000 share, rather than the IP address of the NAS 8000 if
possible. The use of the UNC name is usually easier to maintain if the IP address should change. Care
must be taken to ensure that the drive mapping is 'permanent' on the Windows server, and that all security
associated with the drive mapping is kept current.
preferred method for specifying access points/shares
UNC Names are preferred over mapping a network drive for several reasons. Mapped network drives
are highly "transient" in nature. That is, the drive mappings only exist when that specific user is logged in
on the Windows server. For example, the Windows server system administrator administers a network,
including several NAS 8000's. The administrator maps the T: drive to an administrative share on NAS_1,
and U: to a different administrative share on NAS_2. The administrator also creates an administrative
account for the SQL DBA on the Windows server. This account exists on the Windows server (just like the
administrator's account), but has slightly different privileges. The SQL DBA logs on the console and sets up
the environment such that the T: drive points to a share on NAS_2 for a point of sales database system, and
U: points to a share on NAS_1 for an E-Business system. The DBA creates the databases using the mapped
network drive letters and populates them with initial data. Clients are setup to access these databases
through the SQL Server on the Windows server. Unfortunately, because the SQL DBA setup everything to
depend on mapped network drives, once the SQL DBA logs off the Windows server, those drive mappings
no longer exist and all database accesses fail. It does not matter what user owns the SQL service - the
Windows server will attempt to resolve mapped network drives by the current user that is active at the
console. If there is not a user that is logged in on the console, then there are no valid mapped network
drives. If the Windows server were to log in to the console, the network mapped drives would exist - but
they are not pointing to the same shares as the SQL DBA's account. For reasons like this, it is strongly
recommended that using mapped network drives as a naming method be avoided. If it is not possible to
use UNC names, then the Windows server should be dedicated for sole use by SQL server and the
'administrator' account should be used to administer the Windows server and SQL Server with no other
account being allowed to log in at the console. Also, it is crucial that the 'administrator' account always be
active/logged in, or the mapped network drives will no longer exist and the SQL Server service will not be
able to find/access the NAS 8000.
creating new database(s)
A new database can be created either through the use of the SQL Enterprise Management Tools, or through
the use of command file(s) and SQL scripts. The data file(s), and transaction log(s) can reside solely on the
NAS 8000, or can be mixed on the NAS 8000 and on the SQL Server. Mixing the location of the data
files between the NAS 8000 and the SQL Server also means that other aspects of database performance
and management can be taken into account when designing and implementing the database. For
example, it may be advantageous to have some filegroups remain on the SQL Server while all of the other
filegroups are placed on the NAS 8000. Please refer to the section on NVRAM for important information
regarding placing the database data/log files on both the NAS 8000 and the SQL Serve. One benefit of
using a SQL script for creating a database is the same exact database infrastructure can be created on
other SQL Server systems (such as a development or secondary site) quickly and easily through executing
the same SQL script. Please refer to the constraints section for important information regarding the use of HP
Virus Pro Real Time Protection on a file volume with an active SQL Server database.
8