Since recently upgrading from Opennms 23.x to 24.0.0 (via fresh install) we started to notice that old VMs which had been deleted were still present in OpenNMS and alerting as down.
Our setup has just one requisition which is maintained by the DNS Importer which does a zone transfer every 15 minutes and updates $OPENNMS_HOME/imports/pending/req.xml - This is still happening and the data in req.xml is accurate, however nodes which no longer exist in this file used to be automatically deleted and now they are not.
The provisiond log shows the following error which I believe indicates that the delete is attempted but fails due to newly added foreigh-key constraints:
I believe I have tracked the introduction of this constraint to a change made to `opennms-base-assembly/src/main/filtered/etc/create.sql` as part of a huge commit for https://issues.opennms.org/browse/NMS-9273 (Commit ID: dbd1addf0a36ad8d9cb9c6f2888a0d185e268297)
There was also a subsequent issue and PR to update the Liquibase changelog to reflect this in: https://issues.opennms.org/browse/HZN-1499
I believe the way forward will be to update the SQL used by the importer for deleting/updating notes. If I have any success pinning this down and testing a fix I will submit a PR.