Aerospike Loader

Aerospike Data Loader can help in migrating data from any other database to Aerospike. User can dump data from different databases in .DSV format and use this tool to parse and load them in Aerospike server. User need to provide .DSV data files to load and aerospike schema file in JSON format. It parse those .DSV files and load data in Aerospike Server according to given schema in schema files.

• Prerequisites
• Installation
• Dependencies
• Usage
  • Options
  • Config file format
  • Data file format
• Examples
  • Demo example
  • Detailed examples
• Release Notes

Prerequisites

• Java 1.8 or greater
• Maven 3.0 or greater

Installation

• Source code is available on github:

    $ git clone https://github.com/aerospike/aerospike-loader.git
  

• Then build the utility by running following:

    $ cd aerospike-loader
    $ ./build
  

Dependencies

Following dependencies are downloaded automatically:

• Aerospike Java client 7.2.2 or greater
• Apache commons cli 1.7.0
• Log4j 2.22.1
• Junit 4.13.1
• Json-simple 1.1.1

Usage

If you downloaded the jar from the releases page. Use

    $ java -cp aerospike-load-*-jar-with-dependencies.jar com.aerospike.load.AerospikeLoad <options> <data file name(s)/directory>

If you downloaded the source. Use run_loader script along with options and data files.

    $ ./run_loader <options> <data file name(s)/directory>

"data file name(s)/directory" can either be space delimited files or a directory name containing data files. See "Data Files" section for more details.

For available options and their descriptions run with asloader's --usage option.

    $ java -cp aerospike-load-*-jar-with-dependencies.jar com.aerospike.load.AerospikeLoad --usage
    $ ./run_loader --usage

For more details, refer to Options.

Some extra info about internal working:

• There are 2 types of threads:
  • reader threads (reads CSV files) (The number of reader threads = either number of CPUs or number of files in the directory, whichever one is lower.)
  • writer threads (writes to the cluster) (The number of writer threads = number of CPUs * 5 (5 is scaleFactor))

Sample usage of common options:

    $ ./run_loader -h nodex -p 3000 -n test -T 3000 -e 2592000 -ec 100 -tz PST -wa update -c ~/pathto/config.json datafiles/

Where:

Server IP:                                  nodex (-h)
Port:                                       3000 (-p)
Namespace:                                  test (-n) 
Write Operation Timeout (in milliseconds):  3000 (-T)
Write Error Threshold:                      100 (-ec)
Record Expiration:                          2592000 (-e)
Timezone:                                   PST (-tz)
Write Action:                               update (-wa) 
Data Mapping:                               ~/pathto/config.json (-c)
Data Files:                                 datafiles/

Demo example

Example directory contains two files: allDatatype.json and data.csv. Run the following command to load data from data file data.csv.

./run_loader -h localhost -c example/alldatatype.json example/alldatatype.dsv

For more examples, see the examples/README.md.