You should be able to notice the logging level in namenode.log has been updated, without restarting the service. After finishing your diagnostics you can reset the logging level back to INFO
Reset to INFO
$ hadoop daemonlog -setlevel {namenode_host}:50070BlockStateChange INFO