bireme

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

(★ 138)

Bireme is an incremental synchronization tool for the Greenplum / HashData data warehouse

bireme 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# bireme

[![Build Status](https://travis-ci.org/HashDataInc/bireme.svg)](https://travis-ci.org/HashDataInc/bireme)

[中文文档](README_zh-cn.md)

[Getting Started Guide](docs/bireme_guide.md)

Bireme is an incremental synchronization tool for the Greenplum / HashData data warehouse. It currently supports MySQL, PostgreSQL and MongoDB data sources.

[Greenplum](http://greenplum.org/) is an advanced, fully functional open source data warehouse that provides powerful and fast analysis of the amount of petabyte data. It is uniquely oriented for large data analysis and is supported by the world's most advanced cost-based query optimizer. It can provide high query performance over large amounts of data.

[HashData](http://www.hashdata.cn) is a flexible cloud data warehouses built based on Greenplum.

Bireme uses DELETE + COPY to synchronize the modification records of the data source to Greenplum / HashData. This mode is faster and better than INSERT + UPDATE + DELETE.

Features and Constraints:

* Using small batch loading to enhance the performance of data synchronization. The default load delay time is 10 seconds.
* All tables must have primary keys in the target database.

## 1.1 Data Flow

![data_flow](https://raw.githubusercontent.com/HashDataInc/bireme/master/docs/data_flow.png)

Bireme supports synchronization work of multiple data sources. It can simultaneously read records from multiple data sources in parallel, and load records to the target database.

## 1.2 Data Source

### 1.2.1 Maxwell + Kafka

Maxwell + Kafka is a data source type that bireme currently supports. The structure is as follows:

![maxwell](https://raw.githubusercontent.com/HashDataInc/bireme/master/docs/maxwell.png)

* [Maxwell](http://maxwells-daemon.io/) is an application that reads MySQL binlogs and writes row updates to Kafka as JSON.

### 1.2.2 Debezium + Kafka

Debezium + Kafka is another data source type that bireme currently supports. The structure is as follows:

![debezium](https://raw.githubusercontent.com/HashDataInc/bireme/master/docs/debezium.png)

* [Debezium](http://debezium.io/) is a distributed platform that turns your existing databases into event streams, so that  applications can see and respond immediately to each row-level change in the databases.

## 1.3 How does bireme work

Bireme reads records from the data source, delivers them into separate pipelines. In each pipeline, bireme converts them into internal format and caches them. When the cached records reaches a certain amount, they are merged into a task. Each task contains two collections, *delete*  collection and *insert* collection. It finally updates the records to the target database.

Each data source may have several pipelines. For maxwell, each Kafka partition corresponds to a pipeline and for debezium, each Kafka topic corresponds to a pipeline.

![bireme](https://raw.githubusercontent.com/HashDataInc/bireme/master/docs/bireme.png)

The following picture depicts how change data is processed in a pipeline.

![pipeline](https://raw.githubusercontent.com/HashDataInc/bireme/master/docs/pipeline.png)

## 1.4 Introduction to configuration files

The configuration files consist of two parts:

* Basic configuration file: The default is **config.properties**, which contains the basic configuration of bireme.
* Table mapping file: **\<source_name\>.properties**. Each data source corresponds to a file, which specifies the table to be synchronized and the corresponding table in the target database. \<Source_name\> is specified in the config.properties file.

### 1.4.1 config.properties

**Required parameters**

**Note:** The data source name is just a symbol for convinence. It can be modified as needed.

**Parameters for Maxwell data source**

**Parameters for Debezium data source**

**Other parameters**

### 1.4.2 \<source_name\>.properties

In the configuration file for each data source, specify the table which the data source includes, and the corresponding table in the target database.

```
<OriginTable_1> = <MappedTable_1>
<OriginTable_2> = <MappedTable_2>
...
```

## 1.5 Monitoring

**HTTP Server**

Bireme starts a light HTTP server for acquiring current Load State.

When the HTTP server is started the following endpoints are exposed:

The result is organized in JSON format. Using parameter *pretty* will print the user-friendly result.

**Example**

The following is an example of Load State:

```
{
  "source_name": "XXX",
  "type": "XXX"
  "pipelines": [
    {
      "name": "XXXXXX",
      "latest": "yyyy-MM-ddTHH:mm:ss.SSSZ",
      "delay": XX.XXX,
      "state": "XXXXX"
    },
    {
      "name": "XXXXXX",
      "latest": "yyyy-MM-ddTHH:mm:ss.SSSZ",
      "delay": XX.XXX,
      "state": "XXXXX"
    },
  ]
}
```

* *source_name* is the name of queried data source, as designated in the configuration file.
* *type* is the type of data source.
* *pipelines* is an array, every element in which corresponds to a pipeline. (Every data source may have several separate pipelines.)
 - *name* is the pipeline's name.
 - *latest* is produce time of latest change data that have been successfully loaded to hashdata.
 - *delay* is the time period for change data from entering bireme to being committed to data source.
 - *state* is the pipeline's state.

## 1.6 Reference

[Maxwell Reference](http://maxwells-daemon.io/)  
[Debezium Reference](http://debezium.io/)  
[Kafka Reference](http://kafka.apache.org/)

bireme

bireme 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

bireme

bireme 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README