During the downgrade process, running jobs should continue to make progress without restarting.

If there is an unforeseen error during upgrade, run the following command to restore data:

echo "snapshot_restore '20140917'" | hbase shell

Downgrade each HBase Master one by one. Shut down the HBase Master process, switch to the previous version, and start the process:

su - hbase -c "/usr/hdp/current/hbase-master/bin/hbase-daemon.sh stop master"

hdp-select set hbase-master 2.2.0.0-2041

su - hbase -c "/usr/hdp/current/hbase-master/bin/hbase-daemon.sh start master"

Continue with the next downgrade step.