If Provisiond initiates a scan during an SNMP outage, or after a device has been taken offline, the scan will be aborted with a Warning event:

(uei.opennms.org/internal/provisiond/nodeScanAborted)

When attempting to add or remove surveillance categories, for example to move a node from Production to Terminated, the new categories will never take effect unless SNMP starts responding, or the SNMP service is manually deleted from the primary IP interface.

To access graphs through the webgui or historic measurements through the REST API, decommissioned nodes will have to remain in the original requisition for as long as the historic data remains. (Unfortunately, more often than not the devices will be taken offline before the device has been removed from Production, so polling and data collection attempts continue until someone manually deletes the SNMP service, allowing the node scan to complete.)

Because of this, it would be useful to allow Provisiond to proceed with the scan in order to synchronize changes made to the requisition, even if the provisioning will only be partial.