For some systems, memory-related settings that we discussed for Hadoop are common to HBase. So, let's discuss HBase-specific optimization in brief. We will also look at component-wise HBase optimization.
Let's start with Hadoop.
Add the following parameter to the
<property> <name>dfs.support.append</name> <value>true</value> </property>
The preceding code will enable sync on HDFS, which is essential for durable HBase data synchronization and durability. After making this change, we need to restart the cluster.
The following code decides the number of open files on DataNode and should be kept high as HBase keeps a lot of files open. This must be kept to
4096 or above, according to the size ...