When you install Microsoft SQL Server on a cluster, a
specific set of resources that have dependencies on other resources is created.
Do not modify this default dependency tree, except for the changes that are
listed in this article.
SQL Server and SQL Server Agent should not
depend on any resources, other than the resources that are listed in this
article. Microsoft recommends that you do not make other resources dependent on
SQL Server or on SQL Server Agent. If you must have a dependency, follow the
guidelines that are provided in this article.
The following is an example of a default dependency tree for
SQL Server 7.0:
< SQL_Server_Name> SQL Server Agent 7.0 { SQL Server Agent 7.0 }
+(1)-----Depends On-> < SQL_Server_Name> SQL Server 7.0 { SQL Server 7.0 }
+(2)-----Depends On-> Disk Z: { Physical Disk }
+(2)-----Depends On-> < SQL_Server_Name> VServer { Generic Service }
+(3)-----Depends On-> < SQL_Server_Name> Network Name { Network Name }
+(4)-----Depends On-> < SQL_Server_Name> IP Address { IP Address } The default dependency tree for SQL Server 7.0 has the following
implications:
- The SQL Server Agent resource depends on the SQL Server
resource.
- The SQL Server 2005 resource depends on the SQL network name resource and the physical disk resources that hold its data. When mount points are being used together with the physical disks, each mount point must appear as a cluster resource. Additionally, each mount point must also be added as a SQL Server dependency. If only the root physical disks dependency is added and the mount points are not added, database corruption will occur on failover. Database corruption may also occur when SQL Server is restarted without failing over.
- The SQL VServer resource depends on the SQL network name
resource.
- The SQL network name resource depends on the SQL IP address
resource.
- The SQL IP address resource and the physical disk resources
do not depend on any resources.
The following is an example of a default dependency tree for SQL
Server 2000:
SQL Server Agent { SQL Server Agent }
SQL Server Fulltext { Microsoft Search Service Instance }
+(1)-----Depends On-> SQL Server { Instance Name for named instances }
+(2)-----Depends On-> SQL network name (Virtual Server Name) { Network Name }
+(3)-----Depends On-> SQL IP address1 (Virtual Server Name) { IP address }
+(2)-----Depends On-> Disk Z: { Physical Disk }The default dependency tree for SQL Server 2000 has the following
implications:
- The SQL Server Agent resource depends on the SQL Server
resource, and the SQL Server Fulltext resource depends on the SQL Server
resource.
- The SQL Server resource depends on the SQL network name
resource and the physical disk resources that hold its data.
- The SQL network name resource depends on the SQL IP address
resource.
- The SQL IP address resource and the physical disk resources
do not depend on any resources.
If you modify the default dependency tree, only make the
following changes:
- You can add physical disk resources to the SQL Server
resource so that SQL Server can use additional drives.
- When you add OLAP (Analysis Services) installations, see
the following article in the Microsoft Knowledge Base:
308023
(http://support.microsoft.com/kb/308023/
)
How to cluster SQL Server 2000 Analysis Services in Windows 2000 and Windows Server 2003
Important If you add any other resources to the SQL Server group, those
resources must always have their own unique SQL network name resources and
their own SQL IP address resources. Do not use the existing SQL network name
resources and SQL IP address resources for anything other than SQL Server. If
SQL Server resources are shared with other resources, the following problems
may occur:
- Outages that are not expected may occur.
- Service pack installations may not be
successful.
- The SQL Server Setup program may not be successful. If this
problem occurs, you cannot install additional instances of SQL Server or
perform routine maintenance.
Consider these additional issues:
- FTP with SQL Server replication: For instances of SQL
Server that use FTP with SQL Server replication, your FTP service must use one
of the same physical disks as the installation of SQL Server that is set up to
use the FTP service.
- SQL Server resource dependencies: If you add a resource to
a SQL Server group and you have a dependency on the SQL Server resource to make
sure that SQL Server is available, Microsoft recommends that you add a
dependency on the SQL Server Agent resource, instead of adding a dependency on
the SQL Server resource. To make sure that the computer that is running SQL
Server remains highly available, configure the SQL Server Agent resource so
that it does not affect the SQL Server group if the SQL Server Agent resource
fails.
- File shares and printer resources: When you install File
Share resources or Printer cluster resources, they should not be put on the
same physical disk resources as the computer that is running SQL Server. If
they are put on the same physical disk resources, you may experience
performance degradation and loss of service to the computer that is running SQL
Server.
- MS DTC considerations: If you install Microsoft Distributed
Transaction Coordinator (MS DTC) in your SQL Server group and you have other
resources that are dependent on MS DTC, MS DTC will not be available if this
group is offline or during a failover. By default, SQL Server 7.0 requires that
the MS DTC resource be put in a SQL Server group when you are running the SQL
Server Failover Wizard on new installations. If MS DTC has already been
clustered, SQL Server 7.0 uses the existing location and resources but SQL
Server 2000 does not. Microsoft recommends that you put MS DTC in its own group
with its own physical disk resource, if it is possible.
For SQL Server 2005, see the "How to: Add Dependencies to a SQL Server 2005 Resource" topic in SQL Server 2005 Books Online.
Back to the top
