Resource Pools in SCOM 2012 are an advancement over SCOM 2007, where a resource pool can be used to host instances, that have targeted workflows, and make them highly available. This allowed the “All Management Servers Resource Pool” to host the instances that the RMS used to run in SCOM 2007. This allowed for all management servers, in the AMSRP, to automatically load balance the old RMS workflows, across all management servers.
This also is used for thing like the Notifications Resource pool, which hosts two instances (or Top Level Managed Entities) which are the Pool object itself, and the “Alert Notification Subscription Server” which have many monitoring workflows target it to monitor the notification process health.
Well, we can also write workflows and target resource pools. We might do this if we want a workflow to run on the management servers, but be highly available.
In this example, I will take a VERY simple script that does nothing but log an event, and target the All Management Servers Resource Pool.
First, here is my PowerShell script:
$api = new-object -comObject 'MOM.ScriptAPI' $api.LogScriptEvent("momscriptevent.ps1",9999,0,"this is a test event")
This script simply loads the MOM.ScriptAPI which is necessary to perform specific SCOM actions in script, such as logging events to the SCOM event lot, creating property bags, submitting discovery data, etc.
Then, it logs an informational event for the script in the SCOM event log wherever it is running.
Next up – write my rule to run the script.
We cannot use the SCOM 2007R2 Authoring Console to write this rule, as we need to target the Resource Pool object which SCOM 2007R2 does not understand, nor can it reference. If you are most familiar with authoring in that tool, and you really want to use that SCOM 2007R2 Authoring Console, you can do that, and just target something else, like “Windows Server Operating System” and then change the class later in an XML editor.
Here is my manifest section. Note – I need to reference the SCOM 2012 versions of these MP’s since this MP will not work on SCOM 2007:
<Manifest> <Identity> <ID>Target.ResourcePool.Example</ID> <Version>1.0.0.1</Version> </Identity> <Name>Target.ResourcePool.Example</Name> <References> <Reference Alias="SC"> <ID>Microsoft.SystemCenter.Library</ID> <Version>7.0.8427.0</Version> <PublicKeyToken>31bf3856ad364e35</PublicKeyToken> </Reference> <Reference Alias="Windows"> <ID>Microsoft.Windows.Library</ID> <Version>7.5.8500.0</Version> <PublicKeyToken>31bf3856ad364e35</PublicKeyToken> </Reference> <Reference Alias="Health"> <ID>System.Health.Library</ID> <Version>7.0.8427.0</Version> <PublicKeyToken>31bf3856ad364e35</PublicKeyToken> </Reference> <Reference Alias="System"> <ID>System.Library</ID> <Version>7.5.8500.0</Version> <PublicKeyToken>31bf3856ad364e35</PublicKeyToken> </Reference> </References> </Manifest>
Next, my simple rule. Notice – I target the AMSRP class, I add a simple scheduler module to run this workflows every 30 seconds, and I have a simple write action based on Microsoft.Windows.PowerShellWriteAction module.
<Monitoring> <Rules> <Rule ID="Target.ResourcePool.Example.RunSampleScriptRule" Enabled="true" Target="SC!Microsoft.SystemCenter.AllManagementServersPool" ConfirmDelivery="true" Remotable="true" Priority="Normal" DiscardLevel="100"> <Category>Custom</Category> <DataSources> <DataSource ID="SchedDS" TypeID="System!System.SimpleScheduler"> <IntervalSeconds>30</IntervalSeconds> <SyncTime></SyncTime> </DataSource> </DataSources> <WriteActions> <WriteAction ID="PoshWA" TypeID="Windows!Microsoft.Windows.PowerShellWriteAction"> <ScriptName>momscriptevent.ps1</ScriptName> <ScriptBody><![CDATA[ $api = new-object -comObject 'MOM.ScriptAPI' $api.LogScriptEvent("momscriptevent.ps1",9999,0,"this is a test event") ]]></ScriptBody> <TimeoutSeconds>30</TimeoutSeconds> </WriteAction> </WriteActions> </Rule> </Rules> </Monitoring>
That’s it! I will post my full XML as a sample attached to this article.
Now, when I import this MP, ONE of my management servers should start running this workflow. It will be whichever MS is hosting the AMSRP class at that time. This could change as loads are reshuffled, or as management servers are taken down for maintenance.
I have three management servers, SCOM01, SCOM02, and SCOM03. I can see this workflow is running happily on SCOM02:
I will stop the health service on SCOM02, or shut the OS down.
The last event I got from the test script was at 9:09:56 AM.
What happens now, is the other management servers are waiting for a heartbeat failure threshold to take a vote, and evict SCOM02 from the pool. The SCOM database is also a “default observer” and plays a role in the voting process.
At 9:12:36 AM, I start to see the pool manager events coming in, showing that they other management servers are redistributing the workflows. My 9999 event is now being created on SCOM03, with my first event showing up at 9:12:55 AM, or about 3 minutes after SCOM02 went down.
My sample XML is provided below.
Quick download: https://github.com/thekevinholman/Target.ResourcePool.Example
Hello Kevin,
What If I remove MS (which has many agents reporting to it ) from AMSRP; Will workflows from agents pointing to that MS still get submit to AMSRP and run? will agents reporting to that MS lose the fail-over capability if i remove MS from AMSRP?
Please suggest
There is ZERO relationship between agent assignment and resource pool membership. Zip, Nada, None.
Agents do not “submit to AMSRP”. There is no relationship.
Agent failover has nothing to do with the AMSRP.
The AMSRP simply replaced the RMS, by creating a pool for workflows that were previously all loaded on the RMS. That’s it.
Hi Kevin,
I want to set up a new Resource Pool for specific monitoring tasks, such as URL monitoring. I don’t want to target the AMSRP. When I create the RP, the name is set to a GUID. Is there a way to make this a useful name? If not, will the GUID always be the same on any Management Group?
Regards,
Gerald Versluis
Why is the name set to a GUID? How are you creating the resource pool?
only way I can think of doing that by ‘accident’ is have two language settings partially filled in the MP and neither is set to default. ie en-au and en-us. scom will use one of them and the pool name could exist in the ‘other’ unused one resulting in the guid based naming. This is easily corrected by exporting the MP copying missed items and setting a default on one of them. we have this from time to time with our older MP’s which were first created in scom 2007 as they never tagged this right, longer story why.
Hi Kevin,
Do you still have the example xml file available please?
The file is a dead link.
Thank you,
Jean
Fixed link
Hi Kevin, is there a way to target a custom resource pool? We have a custom discovery created and want to limit which management servers we are installing the additional PowerShell modules on (KeyVault, Accounts, CosmosDB).
Hi,
Recently when I had to target some rules to a resource pool that is created from Console I have used as a target “Microsoft.SystemCenter.ManagementServiceRuntimePool” and disabled the rule. After that I create MP with override to enable it for the pool I need. I guess you can define a resource pool in a Management Pack and use it as a target but I have never done it. And keep in mind that any rule/discovery targeted to a resource pool will run only on one member of the pool.
yes creating a new RP in a MP will allow it to be the target of a rule/monitor with no messing around. its actually really easy.
you create the class type with a base of SystemCenter!Microsoft.SystemCenter.ManagementServicePool make sure to add the language stuff for it so its a readable name and away you go (where system center is the alias for Microsoft.SystemCenter.Library)
ie: