Some time ago I posted an article about gateway servers, Ops Mgr R2 and Server 2008. This week I have extended that scenario by adding two more management servers and one more gateway server.

To make this scenario work I had to configure both my gateway servers to use corp-MS03 and corp-MS02 as primary and secondary management servers. You can do that with operations manager command shell.

Log on to a machine with command shell installed, for example a workstation. Run the following script (save it as a .ps1 script file)

$primaryMS = Get-ManagementServer | where {$_.Name -eq 'corp-ms02.corp.contoso.local'}
$failoverMS = Get-ManagementServer | where {$_.Name -eq 'corp-r2.corp.contoso.local'}
$gatewayMS = Get-GatewayManagementServer | where {$_.Name -eq 'dmz01'}
Set-ManagementServer -GatewayManagementServer: $gatewayMS -primarymanagementserver: $primaryMS -FailoverServer: $failoverMS

That script configure the gateway server, DMZ01, to use corp-MS02 as primary management server and corp-R2 as failover management server. When you run this command you have to be aware that when the current management server gets this update it will stop accepting connections from your gateway server (if it is not primary or failover management server in the new configuration). That could result in a orphaned gateway server, the gateway server is not allowed to communicate with any management server. If this happends you will have to logon to your gateway server, check HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Operations Manager\3.0\Server Management Group\<Management Group Name>\Parent Health Services\0 and look for the authenticationName string. You must then run the set-managementserver cmdlet again and change the management server back to the one in the register. To make sure you don´t end up with orphaned gateway servers use your current management server as failover server in your script, then the gateway server can receive new configuration from the new failover server (current management server) and when you have confirm the configuration change, you can update the failover server for the gateway server.

To verify that your gateway server has received the new configuration you can look in the OpsMgrConnector.Config.XML file on your Gateway server, search for Parents in the file. Default location is C:\Program Files\System Center operations Manager 2007\Health Service State\Connector Configuration Cache\.

You can also run the following command shell commands to see he primary and failover server for a gateway server

Get-GatewayManagementServer | where {$_.Name -like 'DMZ01'} | Get-PrimaryManagementServer
Get-GatewayManagementServer | where {$_.Name -like 'DMZ01'} | Get-FailoverManagementServer

If everything looks ok, you could now run the power shell script again, and update the failoverMS to the correct failover management server.

To configure the agent to communicate with the two gateway servers I ran the following two scripts. corp-r2 was the primary/only management server for the agent when I started.

$primaryMS = Get-GatewayManagementServer | where {$_.Name -eq 'dmz01'}
$failoverMS = Get-ManagementServer | where {$_.Name -eq 'corp-r2.corp.contoso.local'}
$agent = Get-agent | where {$_.Name -eq 'DMZ-A01'}
Set-ManagementServer -AgentManagedComputer: $agent -PrimaryManagementServer: $primaryMS -FailoverServer: $failoverMS

after I verified that the agent had recived the new configuration I ran

$primaryMS = Get-GatewayManagementServer | where {$_.Name -eq 'dmz01'}
$failoverMS = Get-gatewayManagementServer | where {$_.Name -eq 'dmz-02'}
$agent = Get-agent | where {$_.Name -eq 'DMZ-A01'}
Set-ManagementServer -AgentManagedComputer: $agent -PrimaryManagementServer: $primaryMS -FailoverServer: $failoverMS

The result of these commands is that my agent (DMZ-A01) can communicate with both gateway servers. Both my gateway servers can communicate with both my management servers. If I shutdown DMZ01 my agent, DMZ-A01, failover to DMZ-02. If I start DMZ01 again the agent will automatic return to DMZ01. If I shutdown corp-MS02 my gateway server will automatic failover to corp-ms03. If I shutdown one gateway server and one management server the agent will still communicate with the management group.

If you want to see the active node for a cluster group you can do that with a task. Create a task target to “Cluster Node”, then add the following command line configuration

Full path to file: %WINDIR%\system32\cluster.exe
Parameters: $Target/Property[Type="MicrosoftWindowsClusterManagementLibrary6065680!Microsoft.Windows.Cluster.Node"]/ClusterName$ group

You can then run the task from the Cluster Node State view in the cluster MP. The task will list status for all available resource groups.

A couple of days I ago when I was working with a SQL cluster I found something interesting. I had problem with a SQL cluster, Operations Manager 2007 R2 only found the first network name in my cluster group, not the one for my SQL cluster. After some time I found the following text in the SQL management pack guide:

Having a SQL Server cluster resource group that contains more than one network name resource might mean that the clustered SQL Server resource is not monitored. For more information, see Knowledge Base article 919594.

…and KB 919594 tells you

To resolve this issue, configure existing resources in the cluster group so that System Center Operations Manager or MOM detects the virtual server, or create and configure the necessary resources. For example, you may have to create a generic application resource that does nothing. This generic application resource will enable System Center Operations Manager or MOM to discover the virtual server on which the resource is hosted.

Note If more than one Network Name resource is configured in a group, System Center Operations Manager or MOM uses the first network name that it detects as the name of the virtual server..

As soon as I had re-organized my cluster resources for SQL into a new cluster group Operations Manager 2007 R2 found the other cluster resource group and the SQL running on it. It also discovered all the SQL components for example databases and logical disks.

There is a KB article about this, KB959865, that shows you another solution, where you can do a override and enable “Multiple Servers Discovery” within the cluster MP.