locked
SQL DB Engine Not Being Discovered RRS feed

  • General discussion

  • Hi,

    Just wondered if anyone has experienced difficulty getting SQL DB Engines monitored when they are hosted on a cluster.  I have several SQL clusters and the cluster nodes and virtual server appear in the SQL Computers view but the there is no Health State icon in the SQL 2005 DB Engine column against the cluster virtual server name entry or the cluster nodes.

    We are running SCOM 2007 R2 with Cumulative Update 1 installed on Windows Server 2008 x64.

    In order to stage the discovery/monitoring of SQL Server to just those SQL Servers and databases that are most critical we have created 2 groups (1 for standalone SQL servers and 1 for clustered SQL servers and populated these with the servers that we want to enable discovery/monitoring via override to the various SQL Object Discoveries/Monitors.

    We have created a Run-As-Account and set it to use a service account that should have the right privileges on our databases as is evident by the fact that standalone databases are being discovered/monitored correctly.

    We then associated this Run As Account to the SQL Discovery Run As Profile and the SQL Monitoring Run As Profile and the Run As Account was distributed to the servers that we had previously added to our custom SQL Servers groups.

    Agent Proxying is enabled on the cluster nodes.

    Do I need to distribute the Run As Account/Profile to the virtual server name as well as the cluster nodes?  Perhaps I need to add the virtual server name entry to my custom SQL Cluster servers group as well.

    We only have one SQL Server 2008 box at this time which is being monitored correctly so its not the SQL DMO component that is missing.

    Tips from anyone who has successfully managed to get clustered SQL Server DB Engines discovered and monitored are greatly appreciated.

    Kind Regards,

    flipa

    Tuesday, June 22, 2010 2:18 PM

All replies

  • Hi,
     
    I set up monitoring on some of our clusters using the Failover Cluster
    Management Pack:
     
    http://www.microsoft.com/downloads/details.aspx?FamilyId=AC7F42F5-33E9-453D-A923-171C8E1E8E55&displaylang=en
     
    We haven't had to monitor SQL clusters yet, but in my experience, it
    will only show one instance of the clustered service (in our case
    Exchange) hosted on the shared cluster name, but you will get alerts for
    all of the nodes if problems arise. You should be able to explore the
    clusters on the Microsoft Windows Cluster\Clusters diagram in the
    monitoring tab of the operations console.
     
    To be on the safe side, you may have the run-as account propagate to the
    node systems and the shared name, but my hunch is that only the nodes
    would be needed.
     
    The Windows Cluster Action Account Run As profile needs to be configured
    with an account that has administrator permissions on the cluster and
    the SQL Server Monitoring and Discovery profiles need accounts with
    sysadmin permissions on the DB instances (see the MP documentation for
    the SQL Server management pack and the Failover Cluster management pack).
     
     
     

    -- Mike Burr
    Tuesday, July 6, 2010 11:37 PM
  • Hi,

    “On a cluster, when there are multiple clustered SQL Server instances, management pack discovery scripts may time out for Database Engines. So Databases is not discovered. For that, use overrides to increase the timeout value. The default timeout value is 300 seconds. I have increased is 400.”

     

    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.  

     


    Nagaraj N
    Thursday, November 11, 2010 1:49 PM
  • The SQL and cluster MP guides recommend using Local System as the agent action account; we had all sorts of "fun" here with SQL/clustering until we gave up and went with local system. It makes life much easier.

    Also, other MPs require Local System as the action account (like the Exchange MPs), so we've standardized here using Local System just to avoid any problems...


    "Fear disturbs your concentration"
    Thursday, November 11, 2010 6:00 PM