HBase optimization for read
10 views
Skip to first unread message
Felipe Faraco
Apr 20, 2014, 2:41:01 PM4/20/14
to hbas...@googlegroups.com
Hi all,
I'm trying to optimize hbase for reads with an async HBaseClient.
Right now I'm getting 80ms latency with an EC2 cluster with 1 m1.large master and 2 m1.large datanodes.
The table design has only 1 column family that holds a value for each row key.
The table currently has around 70million rows.
Any ideas of design , number of splits, number of region servers (default is 100) per datanode, any suggestion at all?
Thanks
I'm trying to optimize hbase for reads with an async HBaseClient.
Right now I'm getting 80ms latency with an EC2 cluster with 1 m1.large master and 2 m1.large datanodes.
The table design has only 1 column family that holds a value for each row key.
The table currently has around 70million rows.
Any ideas of design , number of splits, number of region servers (default is 100) per datanode, any suggestion at all?
Thanks
Reply all
Reply to author
Forward
0 new messages