Looking at the code, I do not see anything other than the manager (master) calling things that would set the table online.
Were the manager(s) started when the replication table was ONLINE? You could try stopping the managers, setting the replication table offline and then restarting the manager(s). If you cannot take the replication table offline with the shell without a manager, you can force the table offline using the zkCLi set command – the zk path is /accumulo/[instance_id]/tables/+rep/state and replace ONLINE with OFFLINE. You could also try: disabling TABLE_REPLICATION property for any tables that you were replicating? Deleting the TABLE_REPLICATION_TARGET if it is set? And any other replication properties that you may have set. Ed Coleman From: Ligade, Shailesh [USA] <ligade_shail...@bah.com> Sent: Thursday, May 12, 2022 9:32 AM To: user@accumulo.apache.org Subject: accumulo 1.10 replication table issues Hello, I used to have a backup cluster and primary was replicating properly. Now i don't need backup cluster, so i deleted replication entries form primary cluster and turned replication table offline. However time to time the replication table comes online and i am seeing entries in master log stating can't find replication targets and replication is utilizing mx 100 threads etc. since this table comes online on its own, i thinking it is affecting some cluster performance 🙁 Question is why replication table is not staying offline? what is the proper way of keeping it offline? What is the best way to remove all entries from this table that even if it comes online, it will not clutter my logs? Thanks -S
