It is probably caused by some of the metrics which reached its threshold e.g. percent memory used > 90%. It can be changed in the "Network Atlas". Open the map and right click on the node and select Properties then select the Metrics tab where you can set metrics and their thresholds.
Finally had an opportunity to look at this further. My ESXi 5.0.0 server has 5 VMs on it. The Orion "ESX Host Details" page shows 6 volumes. When I do a "list resources", the ESX host responds with the 6 volumes. Unfortunately, the ESX host does respond to polling on the extra volume, so Orion considers it in error. The only way to get rid of the constant sub-child error indication for the ESX host was to:
1. Go to Manage Nodes, select the VMware host
2. Expand the device resource list of the node.
3. Find and select the extra volume
4. Delete that volume from the database.
Now, Orion is happy with all the resources for the VMware node.
I believe NPM may have a problem with VMware polling and interpreting resource information.
Thanks for this... solved some issues with volumes that were removed from a handful of my hosts.