Appendix: Known Issues and Troubleshooting

Applies To: Operations Manager 2007, Operations Manager 2007 R2, SQL Server

Free Space widget incorrectly displays data

Issue: Free Space widget displays data provided by 3 performance collection rules:

1. SQL 2012 DBs Used Space (MB) performance collection rule

2. MSSQL 2012: Collect DB Allocated Free Space (MB)

3. MSSQL 2012: Collect DB Used Space (MB)

If intervals for these rules are not synchronized, the widget cannot correctly display the chart and user sees shifted data series or overlaps in the chart.

Resolution: Make sure intervals for the rules are the same. When intervals are synchronized the widget will start displaying chart correctly. Previously collected data will be incorrectly displayed anyway.

SQL Server 2012 Databases Summary Dashboard displays all active alerts if nothing is selected

Issue: Currently the dashboard displays all active alerts if nothing is selected. It’s a standard querying mechanism of the default Alerts widget.

Resolution: Make sure that at least one DB suites the filtering options.

Widgets cannot display performance data if DB name has special symbols like bracers

Issue: Default performance widgets and SQL Server 2012 Databases Summary Dashboard will not display performance data for Databases that have name with bracers. This is a known SCOM issue.

Resolution: There is no known workaround at the moment.

SQL Server 2012 Databases Summary Dashboard may stop refreshing if left open for a long period of time

Issue: SCOM console has a memory leak that becomes noticeable if you have a dashboard with considerable amount of widgets. Because of the memory leak Authoring Console may consume all available memory and cause a situation when widgets stop refreshing.

Resolution: Restart the Authoring Console.

Functions fail if instance name contains Asian characters

Problem: Functions fail to work on SQL Server 2008 if the instance name contains Asian characters. This situation is caused by the System Center Operations Manager 2007 and System Center Operations Manager 2007 SP1 known issue of the SCOM.ScriptAPI object which is used to pass data from discovery script to Health Service at target machine. This object incorrectly converts Variant to BSTR data type. For more information, see the Microsoft TechNet System Center Operations Manager Authoring forum.

Resolution: This issue is resolved in System Center Operations Manager 2007 R2. Users who need to monitor SQL Server with a localized instance name should update to Operations Manager 2007 R2.

10102 events from “Health Service Modules” are being generated on systems with 32-bit instances of SQL Server running on a 64-bit operating system, indicating that “PerfDataSource” could not resolve counters

Problem: On agent-managed systems that have a 32-bit instance of SQL Server installed on a 64-bit operating system, many 10102 error events will be reported into the Operations Manager log, indicating that PerfDataSource could not resolve a number of counters and that the module will be unloaded. These events are immediately followed by an 1103 event from HealthService indicating that one or more rules or one or more monitors failed and that the failed rules or monitors were unloaded.

Resolution: No resolution. Only a subset of monitoring will work for 32-bit installations of SQL Server that are monitored on a 64-bit operating system. This is because the SQL Server processes are 32-bit, the Operations Manager agent processes are 64-bit, and there are limitations in collecting performance data across architectures. These limitations are documented in Knowledge Base article 891238.

11771 events from “Health Service Modules” are reported on systems with SQL Server 2005 or SQL Server 2008 Management Pack running the SQL Server Full Text Search Service Monitor

Problem: Agent-managed systems are running the SQL Server Management Pack, and many 11771 events are reported. For example, see the following 11771 event.

Resolution: The SQL Server Full Text Search Service Monitor expects to monitor the running SQL Server Full Text Search Service. If the SQL Server Full Text Search Service is not running, the management pack generates the 1171 events. To stop receiving the 11771 events, you can disable the SQL Server Full Text Search Service Monitor.

On clustered SQL Server, Management Pack discovery scripts time out for SQL Server Database Engines, SQL Server Analysis Services, and SQL Server Reporting Services

Problem: On a cluster, when there are multiple clustered SQL Server instances, management pack discovery scripts may time out for Database Engines, Analysis Services, and Reporting Services. You might see events similar to the following example.

Resolution: Use overrides to increase the timeout value. To identify a new timeout value, use the following guidelines: (number of Database Engine instances on the physical node multiplied by 25 seconds*) plus (number of Analysis Services instances on the physical node multiplied by 25 seconds*) plus (number of Reporting Services instances on the physical node multiplied by 25 seconds*).

*This number approximates the time it takes for a script to run. The time it takes for a script to run on your system may vary.

For example, for a clustered SQL Server environment where there are 12 instances of Database Engine on each physical node, 10 instance of Analysis Services on each physical node, and 8 instances of Reporting Services on each physical node, the new timeout estimate is based on the calculation (12*25 sec)+(10*25 sec)+(8*25 sec). The default timeout value is 300 seconds. In this case, the new timeout value should be 750 seconds.

To use an override to enter a new timeout value for discovery scripts

1. Open the Operations console, and then click Authoring.

2. In the Authoring pane, in the left navigation area, click Management Pack Objects, and then click Object Discoveries. The Object Discoveries pane displays a list of objects that the management pack discovers.

3. Right-click the discovery that you want to change, click Overrides, click Override the Object Discovery, and then click For all objects of type: Windows Server. The Override Properties dialog box is displayed.

4. Under Override-controlled parameters, select the Override box next to the Timeout Seconds parameter, and then increase the default number of seconds in Override Value by typing in a new number.

5. Under Management Pack, in Select a destination management pack, click the arrow to select a management pack from the list, or click New and follow the instructions in the wizard to create a new management pack for storing overrides and other customizations. For more information about creating a new management pack for customizations, see the Create a New Management Pack for Customizations section.

Data file and log file free space and free space percent collection rules and monitors might return inaccurate size values

Problem: On some agent-managed systems with databases that are two terabytes or larger, or with databases that were upgraded from SQL Server 2000, free space and free space percent collection rules and monitors might return inaccurate values for data file and log file sizes. The following rules and monitors are reported to have been affected for SQL Server 2005, SQL Server 2008, and SQL Server 2012:

Database Size Provider (Optimized)

• Collect Database Free Space (MB)

• Collect Database Free Space (%)

• Collect Database Size (MB)

• Collect Transaction Log Free Space (MB)

• Collect Transaction Log Free Space (%)

• Collect Transaction Log Size (MB)

Database / Log Free Space

• Database Space Free (%)

• Database Space Free (MB)

• Transaction Log Space Free (%)

• Transaction Log Space Free (MB)

Database Percentage Change

• Percentage Change in Database % Used Space

Resolution: If you encounter this problem, run the DBCC Updateusage (Transact-SQL) command against the affected databases as described in SQL Server 2008 Books Online (February 2009).

No other workarounds exist. If the problem persists, contact customer support services or disable the affected rules or monitors.

Rules and monitors that are based on events from the event log do not work reliably on clustered installations of SQL Server

Problem: On clustered installations of SQL Server, rules and monitors that are based on events from the event log do not work reliably. The problem occurs because events for clustered installations of SQL Server are generated to event logs differently based on the version of the operating system.

Resolution: To resolve this issue, the following prerequisites must be met:

1. Ensure that the version of the SQL Server Management pack that is imported is version 6.0.6648.0 or greater.

2. Check the version of the management group:

   • If the management group is running Operations Manager 2007, upgrade to Operations Manager 2007 SP1 and install the update from Knowledge Base article 959865, Issues that are resolved by the Operations Manager Module rollup update for System Center Operations Manager 2007 Service Pack 1.

   • If the management group is running Operations Manager 2007 SP1, install the update from Knowledge Base article 959865, Issues that are resolved by the Operations Manager Module rollup update for System Center Operations Manager 2007 Service Pack 1.

   • If the management group is running Operations Manager 2007 R2, no additional update is required.