Skip to content

Latest commit

 

History

History
166 lines (106 loc) · 4.48 KB

README.md

File metadata and controls

166 lines (106 loc) · 4.48 KB
Raw

Deployment of DingoDB

DingoDB is a distributed multi-modal vector database that combines the characteristics of data lakes and vector databases and can store data of any type and size (Key-Value, PDF, audio, video, etc.). It has real-time low-latency processing capabilities, enables rapid insight and response, and can efficiently perform instant analysis and process multi-modal data. In order to simplify the deployment, this project introduces the deployment of DingoDB using ansible.

1. Cluster Mode

Physical Topology about DingoDB

The roles in the cluster are mainly divided into:

  • Coordinator

Coordinator act as the master of the cluster. It is responsible for the management and scheduler of data replications of DingoDB cluster.

  • Store

Store act as the storage of the cluster, It is responsible for managing the entire storage.

  • Index

Index is a specialized version of Dingo-Store. It not only offers distributed data storage capabilities but also ensures real-time construction of high-dimensional vector data and its indexing. It provides extensive search capabilities in high-dimensional spaces. The system supports vector index formats such as HNSW and IVF.

  • Executor

Executor act as the worker of the cluster. It is responsible for executing the physical execution plan of SQL to scan and compute the data.

  • Proxy

Proxy provides a bridge for index operations, and provides http and grpc interfaces for use by pythonSDK

  • Driver-MySQL/Driver-DIngo

DingoDB uses JDBC driver to perform table-level data operations, such as create, insert, update, delete, etc. Driver Proxy act as the proxy of JDBC Connection.

1.1 Installation prerequisites

  • Version of OS

CentOS 8.x.

  • Repository of Yum works fine

The repository will be used to install basic tools needed by the cluster, such as python3.

  • Ansible Host

A host installed with ansible is required to distribute cluster configuration and related software modules about DingoDB. This machine can also be replaced by one node in DingoDB cluster such as Store-1 or Coordinator-1.

1.2 Deployment Guidelines

In the cluster mode, ansible is selected as the deployment tools. You can use this guide to install a DingoDB cluster.

1.2.1 Install Steps

You can follow this guide to install a dingo cluster:

asciicast

1.2.2 Installation Notes

1. define cluster configuration

Edit the configuration inventory/hosts, use the real host, user, password to replace the item.

[all:vars]
ansible_connection=ssh
#ansible_ssh_user=root
#ansible_ssh_pass=Password@123
ansible_python_interpreter=/usr/bin/python3

[scaling_in_dingo:children]
add_coordinator
add_store

[coordinator]
172.20.3.201 
172.20.3.200
172.20.3.202

[store]
# 172.20.3.201
# 172.20.3.201 store_num=2 
# 172.20.3.201 store_num=2 disk='/home/sd1 /home/sd2'
172.20.3.201
172.20.3.200 
172.20.3.202

[index]
# 172.20.3.201 index_num=2 disk='/home/sd1 /home/sd2'
172.20.3.201
172.20.3.200 
172.20.3.202

[all_nodes:children]
coordinator
store
index

[executor]
172.20.3.201
172.20.3.200
172.20.3.202

[proxy]
172.20.3.201

[executor_nodes:children]
executor
proxy

2. Check Python3 is installed

Check Python3 is installed or not on DingoDB cluster, if Python3 is not installed, We can use ansible to install Python3 using such command.

ansible all_nodes --become -m raw -a "yum install -y python3" -i ansible_hosts

3. Start to install

  • Copy artifacts
1. artifacts/jdk-8u171-linux-x64.tar.gz
2. artifacts/dingo.zip
3. artifacts/dingo-store.tar.gz
  • merge dingo.zip and dingo-store.tar.gz to dingo.tar.gz
 cd artifacts
 bash merge_dingo.sh
  • check whether there are ports in use in the nodes, and if so, modify the port configuration in Configuration group_vars/all/_shared.yml
ansible-playbook playbooks/02_1_check_ports.yml -e "check_port_conflicts=true"
  • Executor ansible script
 ansible-playbook playbook.yml

2. Docker compose mode

Due to the network firewall, for the convenience of developers, DingoDB team no longer provide a unified GitHub docker repository.

2.1 Installation prerequisites

  • docker
  • docker-compose

2.2 Install Steps

asciicast