IBM PVU Calculated Consumption does not become Consume
I have an issue with some VMs returning a Calculated consumption on an IBM PVU license but the PVU Count does not become a Consumed value.
I understand the conditions that need to be met, based on the OnlineHelp "Relationships Between Consumption Fields":
The system-calculated consumption figure (Calculated consumption) becomes the final result (Consumed) for this device when all of the following conditions occur:
- The value of Overridden consumption for the inventory device is zero (or, for IBM PVU licenses, the Calculated consumption value was imported from ILMT, as shown in the Calculated by column, since in this case Overridden consumption has no effect and acts as a zero)
- For a cores-based license (such as IBM PVU), the inventory source has returned a valid Cores count for the inventory device (or a vCPU count for instances hosted in a known cloud service provider).
I can see both conditions being met, there is no overridden value and there are core counts for each VM so I'm a bit at a lost here. One commonality between these VMs is that they're missing a serial number and they're all LPARs although I'm unsure if this would prevent FNMS from reflecting the final Consumed result.
In addition, the Calculated consumption is actually correct. It's calculating VM Core count x PVU per Core (e.g., 4 x 120 = 480 PVU)
Running FNMS 2020 R2, PVU calculated by FlexNet (Internal)
Any ideas appreciated
@wwwyin - By default, IBM PVU is calculated at full capacity at the host level. Therefore, the consumption of PVU points is shown at the host level, and not at the individual VM level, even if you are calculating PVU at the sub-capacity level. Is this what you are describing?
@kclausen Understand what you're referring to but not what I'm describing. I'm questioning why the Calculated Consumption value is not being brought over to the Consumed column which affects my Peak consumption. For example, Calculated Consumption = 480 PVU, then Consumed = 0
@bfaller Good point, however I just checked again and can see the Hosts are known with Active Status.
I'll throw another spanner in the works - when a manual recon is performed (by publisher), the Consumed PVU points corrects itself but after an overnight full recon, it appears to revert back to 0 Consumed. Keeping in mind this only impacts 4 LPARs under this license with inventory from BigFix. The other VMs (SCCM) seem to be working just fine, so maybe a BigFix issue....
@wwwyin - Since your source of inventory is BigFix, I assume that you are using BigFix as the source of truth for your IBM PVU Sub-Capacity Reporting? If that is correct, can you verify the PVU calculations for those VMs within BigFix? We should be replicating the consumption of BigFix within FNMS.