I also experienced the error in a three node cluster with version 2.8.

But the error disappeared after doing nothing for half an hour...

It looks like the IDM-nodes need some time to get fully in sync with the database.

I am using the SQL-server in this configuration.