Are you using any compression?
________________________________ From: Vimal Jain <[email protected]> To: [email protected] Sent: Thursday, October 17, 2013 6:15 AM Subject: Sudden increase in disk utilization Hi, I have configured Hbase in pseudo distributed mode. I have only one table in hbase (event_data). When i run following command it shows that this table has utilized 8.38 GB. Command :- *[hadoop@woody logs]$ $HBASE_HOME/bin/hbase org.jruby.Main ~/dus.rb /hbase/event_data* O/P :- *Warning: $HADOOP_HOME is deprecated. hdfs://192.168.20.30:9000/hbase/event_data 8.38 GB * where dus.rb is *include Java import org.apache.hadoop.util.StringUtils path = ARGV[0] dus = %x[$HADOOP_HOME/bin/hadoop fs -dus #{path}] splited = dus.split byteDesc = StringUtils.byteDesc(splited[1].to_i) puts splited[0] + "\t" + byteDesc* In short hbase command shows that disk utilized is around 9 GB but actual file system usage is 30 GB. However this difference vanishes if i restart my Hbase. I am not able to understand this. Does Hbase keeps lot of files ( in GBs ) open while working ? -- Thanks and Regards, Vimal Jain
