Cluster Troubleshooting
Applicability:
SQL Server 2000: Not Tested
SQL Server 2005: Not Tested
SQL Server 2008: Not Tested
SQL Server 2008R2: Not Tested
SQL Server 2012: Not Tested
Credits:
Author: Unknown
Date: 22 Aug 2008
Description
Again, I'm not too certain of the source of this information, and while I haven't fully tested all of these scenarios, I certainly have performed a lot of cluster troubleshooting over the past 10 years and have touched on many of these areas.
In my experience, there are as many varaiations of cluster errors as there are network setups, so hopefully these will provide a good starting point for investigations rather than a definitive answer.
1. Check for latest (and consistent) SQL Server Service Pack Level.
Ensure all nodes have exactly the same service pack and hotfixes installed.
How to obtain the latest service pack for:
How to determine the SQL Server patch level from the build number (build number found by running SELECT @@VERSION):
Many thanks to Steve Jones at SQL Server Central.com for keeping these build list up to date
2. SQL Server 2005 64-bit installation on a 64-bit cluster node
This may fail with the error message "The drive specified cannot be used for program location".
This problem occurs because some components of SQL Server 2005 64-bit are 32-bit components and
the 32 components installed in the wrong location when you install SQL Server 2005 64-bit to a
root drive on the cluster node. To work around this problem, do not install SQL Server 2005
64-bit to the root folder of a 64-bit cluster node.
3. When you try to upgrade SQL Server 2005 to SQL Server 2005 service pack 1, the cluster setup may fail.
To work around this problem, you should add the SQL Service account to the domain group for the
MSFTESQL (Full-Text Engine) service before you install SQL Server 2005 service pack 1.
4. SQL Server 2005 failover cluster installation on a node may fail with the error message "The drive specified cannot be used for program location".
This problem occurs when you install the SQL Server 2005 failover cluster by using an account
that is not a member of the Administrators group on the other cluster nodes. To resolve this
problem, include the domain account the SQL Server Setup program runs under to the Administrators
group on all cluster nodes.
5. When you create a WMI (Windows Management Instrumentation) event alert by using the sp_add_alert stored procedure in SQL Server 2005 cluster, you may receive error 22022.
This problem occurs when the SQL network name in a cluster contains 15 or more than 15 characters.
To work around this problem, you can rename the SQL network name so than the SQL network name
contains fewer than 15 characters.
6. When you browse a clustering model of SQL Server 2005 Analysis Services in Business Intelligence Development Studio (BIDS), the Cluster Discrimination tab is empty.
This problem occurs when you create a clustering model that contains one or more input columns
with the "Continuous" content type and the training data contains a null value for these columns.
To work around this problem, you can change the content type to "Discrete" instead of "Continuous"
for the input column that contains null training data or eliminate the null training data by
replacing the null values with 0.
7. After renaming the virtual server name of the default instance of SQL Server 2005, you cannot bring the SQL Server 2005 group online because of a failure in the SQL Server resource.
To resolve this problem, you should install the Cumulative Update package 2 for SQL Server 2005
service pack 2.
8. You may find that replication latency is high when you use transactional replication in a multiple-instance failover cluster environment of SQL Server 2005.
In this case, in the Status column for the subscription that synchronizes with the publication
appears the "Performance critical" warning message. To resolve this problem, you should install
the Cumulative Update package 5 for SQL Server 2005 service pack 2.
9. The SQL Server 2005 Setup program does not remove all IP address cluster resources when uninstall a virtual server instance of SQL Server 2005 in a cluster environment.
To resolve this problem, you can remove the remaining IP address cluster resources manually.
10. When you try to connect to a clustered SQL Server 2005 named instance by using the "servername\instancename" syntax, you receive the following error message: "SQL Server does not exist or access denied".
This problem occurs when you connect to a SQL Server 2005 named instance by using TCP/IP sockets
and IPSec policy is enabled on the client domain, but is not enabled on the server domain.
To work around this problem, you can connect to a clustered SQL Server 2005 named instance
by using Named Pipes or enable the IPSec policy on the server domain.
11. When you install a clustered instance of SQL Server 2005, you may receive the following error message: "Remote setup failed".
This problem occurs when other remote sessions are active on the remote nodes. To work around
this problem, you can log off all the remote nodes of the cluster and then restart the Setup
program.
Code
N/A