We are seeing behavior where newly added devices through VNA are coming through with IPs that are not their management IP.
This is causing them to not sync one to one with the existing devices which were already in the system causing duplicates.
Virtual Network Assurance and Performance Management releases r20.2.3 and earlier
The item being discovered has the UniqueID set in the data. It matches an existing item that has that UniqueID, but a different IP address set.
This prevents the application code from properly reconciling them as the same device.
This is resolved via DE479296 which is available in the r20.2.4 and newer Performance Management releases. It's referenced in the following entry in the Fixed Issues documentation.
There is no workaround to resolve the issue in releases r20.2.3 or earlier. Only an upgrade to r20.2.4 or newer would resolve this.
The new code that comes with an upgrade will prevent future instances of duplicates.
The new code does not resolve existing duplicate devices. They will need manual intervention post upgrade to resolve them. To fix these duplicates we'd complete the following:
This will result in the loss of SDN history but retention of SNMP historical data.
Deleting and then re-importing and rebuilding the plugin is an option. Though it's only an option recommended where hundreds and hundreds of devices are present making manual deletion an administrative nightmare.
In all situations ensure appropriate VNA and PC MySql DB backups are created for use in recovery if needed.
Another way to fix this issue:
Stopped wildfly service (service wildfly stop)
Change the /ca/VNA/plugins/ACI Plugin/config/fabrics/controller.pespective
Changed the expression in - name instaceId
From:
- name: instanceId
expression: path( 'attributes.dn' ).?split( '/sys' )[0]
required: true
To
- name: instanceId
expression: path( 'attributes.serial' ) + path( 'attributes.dn' ).?split( '/sys' )[0]
required: true
Started wildfly service (service wildfly start)