VMs that are onboarded or vMotioned are not reflected in Project or Cloud Zone usage metrics in Aria Automation 8.18.1

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

In VMware Aria Automation 8.18.1, certain vSphere virtual machines may not appear in Project or Cloud Zone usage metrics, even though they are visible in the Deployments and Virtual Machines views. This impacts quota enforcement and reporting, and you may observe that the allocated resource data (CPU, Memory, Storage, or Instance counts) for a Project or Cloud Zone is inaccurate.

You may also see a negative integer value for the consumption of resources when polling for resource consumption statistics through the API. A negative value indicates that the Project has been over-allocated and subsequent deployment requests on this Project will fail.

Environment

• Aria Automation 8.18.1 GA, 8.18.1 Patch 1, 8.18.1 Patch 2, 8.18.1 Patch 3

Cause

This typically occurs when managed VMs or disks are missing proper placement links in the underlying database, leading to incorrect resource accounting.

This can happen in the following scenarios:

• Moving VMs (using vMotion) across clusters, across datacenters or across vCenters
• Onboarded VMs into Aria Automation that are now visible in the UI but not counted against project or Cloud Zone quotas.

Resolution

This issue is resolved in 8.18.1 Patch 4, but any VMs that were onboarded or vMotion-ed prior to 8.18.1 Patch 4 will need to be manually remediated as per the instructions in this KB.

If you are using multi-tenancy in Aria Automation and have deployed VMs in VPCs, this KB will need to be used as the fix for the multi-tenancy scenario is targeted for an upcoming patch release of Aria Automation.

Step 1 - Identify the VMs and Disks that have improper placement links and identify the Projects that will be overallocated when the VM links are corrected.

Extract the attachment and run query_corrupted_machines_v5.sh

This will output the following:

1. VMs that have incorrect placement links
2. Disks that have incorrect placement links
3. Projects and Placement Zones that will end up with overallocated (negative) resource usage metrics when the VM and Disks links are corrected.

Example:

Step 2 - Fix overallocated (negative) Project placement zones

Review the Projects that were listed in Step 1 that have overallocated (negative) resource limits. For each of the projects listed in step 1 that shows overallocated placement zones, you must do one of the following to fix the Project’s limits:

• Increase the relevant limit (CPU, GPU, Memory or Storage) in the zone to the total that is mentioned in the output of step 1. OR
• Perform Day-2 Change Project to move resources to a different Project OR
• Free up resources by deleting deployments from this zone

Step 3 - Fix the placement links of the impacted VMs and Disks

Run grpremediate_by_project_v5.sh.

This will fix the placement links of the VMs and Disks.

The script provides options to apply the fix on either a specific project or on all impacted projects. Run grpremediate_by_project_v5.sh -h to get usage.

Example:

Step 4 - Recalculate and update the Project resource metrics

Follow KB 326134 to recalculate and update the resource metrics

Note

Even after upgrade to 8.18.1 P4, moving VMs across datacenters or vCenters can still lead to Project’s limits being overallocated. However, you will now immediately notice this in the Project’s resource usage in the UI.

To fix the Project’s limits, you must do one of the following:

• Increase the relevant limit (CPU, GPU, Memory or Storage) in the zone. OR
• Temporarily set the relevant limit to 0 (unlimited). Then delete the deployments to free up resources and set the limit back to the original value. OR
• Perform Day-2 Change Project to move resources to a different Project

Additional Information