Info: This issue has been solved by accessing and deleting them directly the chirpstack_ns database and for some cases in chirpstack_as as well.
after the delete it is possible to ADD back the device properly.
We’ve just had the same issue on a new V3 instance but with over 1000 devices and we’ve had it occur on older versions of Chirpstack as well. Is there any explanation as to why this happens? Is there a way to re-index the database or something? or do we really have to delete these and re insert them? How do we know it won’t happen again?
If the databases are out of sync, then this would require a manual action to delete the “ghost” records. Please note that this is one of the improvements of ChirpStack v4, where data is no longer separated in two databases.