Skip to content

Docker images for building hadoop3.2, hive 3.1, hbase2.3, presto 0.247, flink1.11.3 on yarn, etc.

Notifications You must be signed in to change notification settings

lawlite19/bigdata-docker-builds

 
 

Repository files navigation

bigdata-docker-builds

Docker images for building hadoop3.2, hive 3.1, hbase2.3, prestodb 0.247, flink1.11.3 on yarn, etc.

Get Started

You can use docker to set up hadoop based big data platform in a few minutes, docker images include Hadoop 3+, HBase 2+, Hive 3+, Kafka 2+, Prestodb 0.247+, Flink 1.11+, ELK 7.9+ ,etc. Integration tests between Hadoop and Hive, Hadoop and HBase, Flink on yarn, Prestodb against Kafka, Elasticsearch, HBase, Hive had been covered. You can used it in development evironment to test your applications, but it's not recommended to use in production.

Sample docker-compose.yml holds here, just download it and get started right now.

Supported components

  • Apache Hadoop 3.2
  • Prestodb 0.247
  • Kafka 2+
  • Hbase 2.2
  • Hive 3.1.2
  • ELK 7.9.1

How to Run

Git clone the repo and run docker-compose.

docker-compose up -d

How to setup docker & docker compose

Install docker ce on centos 7 or above:

1. yum remove docker docker-common docker-selinux docker-engine
2. yum install -y yum-utils device-mapper-persistent-data lvm2
3. yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
4. yum install -y docker-ce
5. systemctl start docker.service
6. systemctl enable docker.service

Install docker-compose:

1. sudo curl -L "https://github.com/docker/compose/releases/download/1.23.2/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose

2. sudo chmod +x /usr/local/bin/docker-compose

3. docker-compose --version

How to visit services in web ui?

Note: you have to add the server ip and services (which defined in docker-compose.yml) to your local hosts firstly. [how to configure hosts](https://www.howtogeek.com/howto/27350/beginner-geek-how-to-edit-your-hosts-file/)

## Produce some data to test HDFS & Hive

 <pre>
   create a table in hive:

   ```hive sql
   create external table test(
    id      int
   ,name    string
   ,skills   array<string>
   ,add     map<String,string>
    )
    row format delimited
    fields terminated by ','
    collection items terminated by '-'
    map keys terminated by ':'
    location '/user/test'
create a txt file with the content below, put it under /data/ directory, such as
/data/example.txt

```
1,spancer,bigdata-ai-devops,changsha:lugu-changsha:chuanggu
2,jack,webdev-microservices,shenzhen:nanshan-usa:la
3,james,android-flutter,beijing:chaoyang
3,james,ios-flutter,beijing:chaoyang
```
load local file data into the external hive table, which we created above.

```
load data local inpath '/tools/example.txt'overwrite into table test; 

```
Check the mr job in yarn web ui :http://resourcemanager:8088/

After the job is done, query in hive client. Alternative, we can query the data in presto client.

```
select * from test;
```

Testing HBase

create a table named as person, with two column families, info and tags.

```
create 'person', 'info','tags'
```

list tables in HBase
```
list
```
describe table person

```
describe 'person'
```

Insert some data into table person with rowkey 1.

```
put 'person','1','info:name','spancer'
put 'person','1','info:age','36'
put 'person', '1', 'tags:skill','bigdata'
```

scan table person
```
scan 'person'
```

get data by rowkey
```
get 'person','1'

```
get data by rowkey and column family or column.

```
get 'person','1'

get 'person','1','info'

get 'person','1','info:name'

How to test prestodb to confirm whether it works?

1. cd the prestodb container: docker-compose exec prestodb bash
2. connect to presto using presto-client: presto --server prestodb:8080 --catalog elasticsearch 
3. query some data: show schemas; show tables; select * from nodes;
4. you can also verify presto status through web ui: http://your-host-ip:9999

Run jobs on the docker platform.

For java developers, I provide some tests over the platform. You can fork it from here, the test project contains flink jobs with set of components, such as kafka, elasticsearch, iceberg, etc.. Source and sink examples are fully inclued.

ToDo

  1. Integration flink 1.12
  2. Integration hive 3.1 (Done)
  3. Integration hbase 2.2
  4. Integration iceberg
  5. Integration alluxio
  6. Integration ozone

About

Docker images for building hadoop3.2, hive 3.1, hbase2.3, presto 0.247, flink1.11.3 on yarn, etc.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Shell 79.5%
  • Dockerfile 18.4%
  • Makefile 2.1%