Data deletion in Discovery is based on aging parameters.
These parameters are available for:
- Directly Discovered Data (a set of which gets generated each scan)
- Inferred model (Hosts, Software Instances, etc).
These are controlled separately. The settings relevant to Host aging have defaults of 10 days and 7 failed scans (both must be satisfied) before Discovery marks the Host as destroyed. See the documentation for more details.
When a host ages out of Discovery
- The continuous CMDB sync sends a delete action to mark as deleted the corresponding CIs in the Discovery CMDB dataset (BMC.ADDM by default)
- The reconciliation jobs in CMDB propagates this change into BMC.ASSET
- If BMC.ADDM has a purge job, its MarkedAsDeleted CIs will be hard deleted (If the related CI in BMC.ASSET is up to date).
- If BMC.ASSET has a purge job, its MarkedAsDeleted CIs will be hard deleted
If continuous CMDB sync is not running when Hosts or Devices are aged out or destroyed via the UI, then the CMDB won't be updated.
The following query returns the list of nodes deletion that were "missed" while the sync was not running
search flags(include_destroyed, exclude_current)
Host, NetworkDevice
where destructionTime(#) > last_cmdb_sync_success
show name, age_count, last_update_success, os, destroyed(#) ,
friendlyTime(destructionTime(#)) as destructionTime,
last_cmdb_sync_success
Any destroyed Hosts returned by this query were not sync'd to the CMDB after they were destroyed. To resolve this, manually sync the Hosts returned by this query so that they are Marked As Deleted in the CMDB.