Deploying OCPM in Multi-Server Environments

Alex Jauch - Architect, Microsoft Private Cloud

 

When deploying OnCommand Plug-In 3.1 for Microsoft, you may have questions about the proper way to install the plug-in for environments where you have more than one SCOM server.  For those environments, it’s important to understand that OCPM only supports an Active-Passive failover mechanism.  This means that while you can have OCPM installed on as many SCOM management servers as you like, only one management server can actively be collecting data at any time.  This behavior is controlled by enabling or disabling the Data ONTAP discovery rule that ships as part of OCPM.

 

Let’s take a more concrete example to illustrate how this would work in practice.  In this case, we will have two datacenters, a primary datacenter (DC A) and a secondary datacenter (DC B).  In support of this configuration, we will create two SCOM management servers SCOM A will be located in DC A and SCOM B will be located in DC B.  As part of this design, we want DC B to take over all functions in of DC A in case of any failure or outage.  In this case, we would install OCPM on both SCOM A and SCOM B.  This ensures that SCOM B can take over as needed.  However, when we override the Data ONTAP discovery rules, we should only target the override to SCOM A.

 

 

Notice in the screen shot above, that we are creating the override “For a specific object of class:  Management Server.”  This is important.  Selecting “For all objects of class:  Management Server” will enable the rule for all management servers which is not what we want.

    

After selecting the management server to target, you will see this screen:

 

 

 

Note that in the screenshot above, I am targeting a specific server (FT-SCOM.flexpod.test in this case) and I am overriding the “enabled” value to true.  Also note that I have created my own management pack to save my overrides (“FlexPod Configuration” in this case).  This is important because if you save these overrides in the default management pack, this will cause a dependency between OCPM and the default management pack and will prevent you from removing OCPM or upgrading to a new version in the future.

 

If for some reason, you want to move discovery and monitoring to another SCOM management server, you simply remove the existing override and add a new one for the secondary server.  If you right click the rule and select “Overrides Summary” you will get the screen below:

 

   

 

Note that I only have one override which is to enable therule for a specific management server.  Click delete on this screen to delete and then create a new rule targeting SCOM B using the procedure above.

    

In the example above, I showed only the Data ONTAP: Discovery Rule, but you should follow the exact same procedure for the Data ONTAP Virtualization: Discovery Rule as well if you are using OCPM to monitor Hyper-V servers.

Comments

Hi Alex,

Is it possible to customize the warning and critical thresholds on aggregates & volumes in this new plugin 3.1?

Thanks

jausch Former NetApp Employee

Yes, all the thresholds can be modified.  However, this is on a systemwide basis, not on a per object basis.