A yes, I want a little PowerCLI and more vOrchestrator, but my next task brings me vCenter Operations Manager. Also a very handy, must have, excellent tool for every vSphere administrator/engineer/architect. Here goes.
vCenter operations manager is the key component of the vCenter Operations Management Suite. It provide a holistics view and deep operational insights into the health, risks and efficiency of the virtual infrastructure and its application workload. It identifies capacity short falls and over-provisioning, with that information you can right size the VM’s, reclaim not used resources and increase consolidation ratio’s. This is all in a proactive performance management solution, with automated root cause analysis and recommended actions to remediate potential bottlenecks.
The operations management suite consists of the following components:
- vCenter Operations Manager, or vCOPS in short.
- vCenter Configuration Manager;
- vCenter Hyperic;
- vCenter Infrastructure Navigator;
- vCenter Chargeback Manager;
This document handles the installation of the key component VCOPs.
vCOPS is a vApp that you import and deploy in the virtualization layer, in this case vSphere ESXi.
The vApp consists of two virtual machines:
- the UI VM. Access to the analytics and access to the administration portal via several web interfaces. Addins/plugins to vSphere client and the vSphere web client let’s you view and manage the environment from a single web control plane.
- the analytics VM. Responsible for collecting data from vCenter and other VMware or third party infrastructure components. Raw metric data is stored in the File System Database (FSDB) and other collected data, objects, relationships, alerts and thresholds are stored in a Postgres DB.
All put together in an architecture model taken from the vmware.com site:
For deployment you will need:
- vCenter to connect to.
- Active Directory user with access to the source vCenter, and access in the infrastructure vCenter to deploy.
- Amount of VM’s to monitor, depending on your license (vSphere edition or vCOPS license) you will need the amount of VM’s licensed and you will need the resource calculated for the amount of VM’s:
Small: less than 1500 VM’s. 4vCPU’s and 16GB vRAM.
Medium: between 1500-3000 VM’s. 8 vCPU’s and 25GB vRAM.
Large: Larger than 3000 VM’s. 16 vCPU’s and 34 GB vRAM.
- vCenter for the infrastructure where your vApp will run.
- Networking information.
- Storage location and capacity, starting with 344 GB Thick provisioned (or 3.8 GB Thin when you want to thin provision).
- Resources, depending on the amount of VM’s.
Download the OVA or OVF and VMDK files from vmware.com. Either import the OVF from the web client or from the vSphere client. When using the vSphere client you will have to create an IP Pool to support the vApp. With the web client this is created upon import and configuration.
Note: when creating the IP Pool in the vSphere client you don’t have to enable the IP Pool and add a IP range. When deploying the vApp you either configure a static IP (recommended) or use a DHCP server present in the IP range.
Go to Inventory – Networking and select the datacenter object you want to deploy vCOPS into. Select the IP Pools tab. Click add.
Fill in the IP Subnet and gateway, DHCP (only when a DHCP server is to be used and present), DNS and Associations. Do not select Enable IP Pool (normally when you want a Transient allocation, but this is not supported with vCOPS). But as I want static, I don’t need the IP Pool to be enabled. Association of the IP pool must be to the portgroup(s) you want you VM’s configured with.
Next up is deploying the vApp. Select deploy OVF Template and browse to your downloaded location. Follow the prompts of the Deploy OVF Template Wizard.
- Accept the EULA.
- Name your vAPP and select an Inventory location in the data center.
- Select your deployment configuration. I select small as the infrastructure is much smaller than 1500 VM’s.
- Select the Cluster or host the vApp needs to run on.
- Select the storage location. Standard the compatibility check is on Thick provisioned, you can change this to go to disk format and then return to Storage location.
- Thick Provisioned Eager Zeroed disk format is recommended, or whatever is the organizations standard is.
- Select Network Mapping to the wanted portgroup.
- Select an IP Allocation Policy, I want Fixed.
- Fill in the required properties and when finshed let it rip.
- Wait for the VM’s to be started
- Check the IP’s in the VM tab’s.
Connect to the vCOPS web service. This can be reached via https://<ip UI VM>/admin. Use the default admin admin combination on first boot.
Configure the following:
- vCenter Hosting server. If FQDN gives error, IP address should be used to import certificate.
- Change password on admin and on root.
- Specify the vCenter to register with (for monitoring). Specify users for registering and collecting information.
- Import data.
Next assign a license via the License management to the vCOPS solution.
When needed set SMTP en SNMP to forward alerts, and add users to the roles to be able to use vCOPS as well.
You can now go to solutions and vCOPS. Let the solution gather some information on the infrastructure before any actions will be done. As more data becomes available, more information is displayed. Depending on the size of the to monitor infrastructure this can take a few minutes to hours.
Check if there are no alerts, both in the admin and in the vCOPS UI.
The dashboard is opened via the vSphere client (after you enabled the plug-in) when you go to Home – solutions and Applications – vCenter Operations Manager. You can also open this in a web client by browsing to https://<UI IP address>/.
You can view the dashboard and details on several point in the infrastructure:
- World shows all connected infrastructures and the combined status.
- vCenter shows that vCenter inventory status.
- From vCenter you can zoom in to the vCenter objects just like in the vSphere infrastructure:
When you are used to the default dashboard and it’s badges you can create custom dashboards to further specialize vCOPS solution to your organization’s needs. The custom interface is reachable via https://<ui IP address>/vcops-custom. Same goes for setting organization specific policies instead to the default policies. This means setting specific threshold values, detection rules, alerts and forecasts and trends.
The standard dashboard is created from the following badges:
- Health. Which in turn is a combination of workload, anomalies and faults. The higher this number, the healthier the environment is. Low number is bad.
- Risk. This is a combination of stress, time remaining and capacity remaining. This is a projected risk in a future condition (near or long term). The lower this number the better, high is bad.
- Efficiency. This is a combination of reclaimable waste and density (consolidation p:v for CPU and memory and VM to host ratio). The higher this number, the better. Low number is bad.
Going from Dashboard to more
Going from the dashboard badges to more specific information, planning, analysis and reporting via the tabs shown on the top pages or by clicking on some of the badges of interest in the dashboard. The active alerts, warnings and informational numbers are shown on the right of the screen.
But for now this concludes this document.