Thursday, April 30, 2015

Part 15: Re-initiate Monitoring Goals Wizard in vRealize Operations 6.0!

One of the post in my vROps 6.0 Tutorial was about understanding the Monitoring Goals in vROps 6.0. I would highly recommend that you read that post before you read this article as this one is a follow up on that post where at the end I mentioned :

"As I mentioned before, any changes to the policy now cannot be done through this wizard. You would have to manually edit the policy now to make changes or you could use my next post where I will tell you a hack which will allow you to re-run this wizard once again."
The above statement means that once you setup a vCenter Adapter to monitor a vCenter Server, you would be asked to define your monitoring goals. These monitoring goals basically translate into the default policy which are then applied on all the objects which are being monitored by that vROps instance. When you read the above mentioned article you would understand how this wizard helps you set the policy for the first time. Unfortunately, you can only run this widget once and set the default policy and to make any subsequent changes to the policy, you would have to manually edit the default policy which could be a cumbersome task if you are not well versed with how policy works!
While you do not have the option to re-run the wizard, there is a hack which was discovered when this question was asked by my colleague Woflgang Stichel who is a Business Solution Architect at VMware. His question was that one of his customer has not selected the right options while responding to the monitoring goals because of which vROps default policy is not monitoring the datastores at all. He wanted to re-run this wizard somehow to select the right options instead of diving into the policies.
Before I share the hack with you, I must mention that this is NOT a supported method and please do it at your own risk :-)


Special thanks to Wolfgang for documenting the process like a PRO... Sharing his notes with you all!!

BACKGROUND

The purpose of this document is to describe, how Configuration Policies in vR Ops are created and how to initiate the Configuration Wizard from scratch, which is usually started just when the product is deployed. Note: This approach does not affect the existing objects. As long as just the Adapter Instances get removed, the objects and its metrics will not be deleted.

The Configuration-Wizards helps to set a bunch of Settings within the Active Policy to define the proper approach how vR Ops calculates capacity and data displayed.

Alternatively the document describes, how to make the changes manually in the active Policy.
So take this document like a „How do I reset my Config-Policies“.

Special thanks to Sunny Dua, who delivered the conceptual idea of this approach.


PROBLEM STATEMENT

Purpose: Understand the raw concept of Configuration Policies and the demand for resetting the active Policy.

Configuration Policies are delivered when vR Ops is installed and when additional Solutions (Management Packs) are installed. They define, which metrics are influencing the Capacity Management, Reports and Alerts. Policies follow an “inheritance model”.

I.E. there could be a nested Policy-Structure like Policy1 -> Policy2 -> Policy3. Let’s assume for HostA Policy3 is active.

When Policy1 states, Network IO should be considered for capacity Management, Policy 2 states no Network IO is considered and Policy 3 is set to inherit this setting, then HostA would not consider NetworkIO for capacity Management.

When the vR Ops is deployed, the Config-Wizard starts and asks for several aspects the Default Policy should be based on (like will Network-Workload be considered for Capacity Management). When finished the Default Policy becomes the active Policy for the whole vSphere Environment. It is possible to define multiple active Policies for different environments afterwards. But this will not be covered in this document.

Depending on your entries into the Config-Wizard, you probably get no Workload Analysis data for Datastore.



The reason for this behavior: In the active Policy Workload for Datastores is disabled, DatastoreIO is not considered for Analysis.


There are 2 Options to solve it:

1.) Re-initiate the Config-Wizard

2.) Make manual changes to the active Policy (Will not discuss this option in this article)

RE-INITIATE THE CONFIG WIZARD

The creation of the active Policies for data of an vCenter is initiated when the Solution is deployed. To re-initiate this process the Adapters of the Solution have to be removed and subsequently added.
In this case the VMware vSphere Solution. Go to Administration -> Solutions



VMware vSphere and Click on Configure

Select vCenter Adapter and remove  the vCenter adapter instance. In this case vc-l- 01a.corp.local. Also remove the vCenter Python Actions Adapter. Then Save Settings and Close.


Next Click Configure for the VMware vSphere Solution



You will now get to the Config-Wizard. Enter the settings for the  vCenter Instance accordingly, select credentials and Click Save Settings. Answer pop-ups for certificates (OK) and already registered vCenters (Yes). Then add the vCenter Python Actions Adapter (if desired). Click Save Settings. Important: Click Next



The Config-Wizard starts and you have the choice to make your changes.
I.E. consider DatastoreIO and/or NetworkIO. In this case I’ll select “Include only Storage”



Then Next and finish.

To double-check the changes are effected, go to Administration->Policies and select active Policies (note actual timestamp in Policyname). Under Details (lower frame) select “Complete settings including inherited” and expand Datastore. You’ll see, Workload is now considered.



In the Analysis-Tab Workload is no longer greyed-out. Also the Policyname has changed (see timestamp).


Hope this will help you re-define your monitoring goals in case you could not do so at the first go!!


Share & Spread the Knowledge!!


No comments:

Post a Comment