The documentation for disabling the peer sounds wrong, I think it applies to option #1, not #2.
>From my memory, disabling/re-enabling the peer did not require any manual >intervention, it just resumed from whatever info was in zk but admittedly, >this was back on 0.94 so things might've changed. -Chien -----Original Message----- From: spats [mailto:[email protected]] Sent: Monday, October 3, 2016 8:18 PM To: [email protected] Subject: Best way to disable & enable HBASE replication If we have to stop hbase replication and re enable after sometime, which one is better way to do 1. stop_replication, later enable by running start_replication and copy missing data for that duration using copyTable command. 2. disable_peer("1") , later enable_peer("1") 2nd option looks better as we don't have to copy missing data manually & data gets replicated automatically when we enable peer. But i see warning here https://www.cloudera.com/documentation/enterprise/5-5-x/topics/cdh_bdr_hbase_replication.html#topic_20_11_5 "If you disable replication, and then later decide to enable it again, you must manually remove the old replication data from ZooKeeper by deleting the contents of the replication queue", don't understand statement clearly. Is it really required to delete znode? I would like to go ahead with option 2, is there any risk? deleting znode is mandatory in that case? Thanks in advance -- View this message in context: http://apache-hbase.679495.n3.nabble.com/Best-way-to-disable-enable-HBASE-replication-tp4083081.html Sent from the HBase User mailing list archive at Nabble.com.
