// Create a connection to the cluster.
HConnection connection = HConnectionManager.createConnection(Configuration);
HTableInterface table = connection.getTable("myTable");
// use table as needed, the table returned is lightweight
table.close();
// use the connection for other access to the cluster
connection.close();
回复 ( 7 )
纠正
准确而言不是被弃用 是不建议client用 都成了developers api
所有client端被弃用的api都有相对应的interface 用那个就好
这样做的好处是 将具体实现隔离 留给开发者更大的开发空间 以后随意抽换更好的实现 而不影响client端代码
说白了就是bridge pattern
I don’t know what version are you using.
From HBase 0.98.x you don’t need to pool HTable object anymore.
Just get Table object from Connection object when you need, all the resource caching and management is done internally by Connection class.
Apache HBase ™ Reference Guide
For example, in multi-threaded environment the standard method can be described like this:
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.client.ConnectionFactory;
import org.apache.hadoop.hbase.client.Connection;
import org.apache.hadoop.hbsae.client.Table;
int main() {
// create a new hbase configuration object. (singleton)
Configuration conf = HBaseConfiguration.create();
// create a new hbase connection object. (singleton)
Connection connection = ConnectionFactory.createConnection(conf);
try {
while (true) {
// create a table instance everytime you need.
// you don’t have to pool this object, because hbase client implementation
// do necessary resource caching & control internally.
Table table = connection.getTable(TableName.valueOf(“table”));
…
table.close();
}
} finally {
// close the hbase connection on application shutdown.
connection.close();
}
}
机器配置太低了。
首先hbase内存至少要12G以上开给JVM, CMS回收，解决JVM hang问题。
台数得根据自己的业务需求压力测试，评估负载。用途不一样，优化方向不一向
发生的问题第一个是memstore, 第二个是compaction. 第三个JVM hang.
memstore, compaction优化后的参数。当然，环境不一样，参数不一定适合，以下给个思路。下面是一个高并发，数据本身小的参数优化。
<!– memstore –>
<property>
<name>hbase.regionserver.global.memstore.upperLimit</name>
<value>0.2</value><!– default:0.4 –>
</property>
<property>
<name>hbase.regionserver.global.memstore.lowerLimit</name>
<value>0.15</value><!– default:0.38 –>
</property>
<property>
<name>hbase.regionserver.optionalcacheflushinterval</name>
<value>3600000</value><!– 1 hour, default: 1 hour –>
</property>
<property>
<name>hbase.hregion.memstore.mslab.chunksize</name>
<value>2097152</value>
</property>
<property>
<name>hbase.hregion.memstore.mslab.max.allocation</name>
<value>1024768</value>
</property>
<property>
<name>hbase.hregion.memstore.flush.size</name>
<value>134217728</value><!– default: 128MB –>
</property>
<!– compaction –>
<property>
<name>hbase.hregion.majorcompaction</name>
<value>86400000</value><!– 1 day, default:7 days –>
</property>
<property>
<name>hbase.hstore.compaction.min</name>
<value>3</value>
</property>
<property>
<name>hbase.hstore.compaction.max</name>
<value>10</value><!– default: 10 –>
</property>
<property>
<name>hbase.hstore.blockingStoreFiles</name>
<value>30</value><!– default: 10 –>
</property>
<property>
<name>hbase.regionserver.thread.compaction.large</name>
<value>2</value><!– default: 1 –>
</property>
98的 API 上有说明可以直接用
1.1.2中HConnectionManager 也被 Deprecated
直接用org.apache.hadoop.hbase.client.Connection;就可以
可以用hconnection代替，线程安全
有人吗？