Is it a good idea to move facts to aerospike from hbase on a daily basis so that queries are faster when compared with impala/hbase? If so, then could you please post an example to load data from HBase to Aerospike?
1 Like
Hi, We don’t have any comparison for performance of query between Aerospike and Hbase. Still you can try loading your data into Aerospike using the following procedure:
- 1 - Dump your Hbase data into CSV format.
- 2 - Use Aerospike-loader to import data into Aerospike.
There is no API where in i can load directly into aerospike from HBase?
Sorry to say there is no direct API/connector to transfer data from Hbase to Aerospike. The only way is to dump into CSV format and use Aerospike-loader to load into Aerospike.
Ok. I was thinking to move facts and big tables to aerospike to improve performance. But it appears like we have to either load data into hdfs and then move to aerospike. I hope we have connectivity from hdfs to aerospike.
Can you give more details on the workflow you envision? As regular batch dump from HBase? Real-time update? What API is most natural?
If you can access your HBase data via MapReduce (which you should), the Aerospike Hadoop connector should help you move data via a simple pass-through MR job to Aerospike.