I. Overview
1. Version
Designer | JAR | S3 Repository Plugin |
---|---|---|
10.0.16 | 2021-05-26 | V1.2.8 |
2. Application scenarios
It is difficult to deploy the HDFS by yourself. Therefore, you need to use a highly available file server.
Users want to use cloud storage file systems that support the S3 protocol as file servers, such as Aliyun OSS, Huawei OBS, and Yamun S3.
3. Function introduction
You can install the "S3 Repository" plugin and use the cloud storage file system that supports the S3 protocol as the file server.
II. Steps
1. Install the plugin
Click to download the plugin: fr-plugin-s3-repository-1.2.8.zip
Designer plugin installation method reference: Plugin Management
Server installation plugin method reference: Server plugin management
2. Cluster environment preparation
Before configuring an S3 file server, prepare a cluster environment. For details, see Cluster Environment Preparation.
3. Config file server
1) Enter the decision-making platform as an administrator, click "Manage > Intelligent Operations > Cluster Configuration", click "Share File Server", and select "S3" for protocol, as shown below:
2) Configure according to the actual situation, as shown in the figure below:
Note 1: "Path" must end with "/" and cannot start with "/".
Note 2: "Path" cannot be a relative path.
3) Click "Test connection and save", indicating that the connection is successful. As shown in the figure below:
Note: For the next steps of configuring a cluster, see Configure and Start Cluster.
III. Precautions
After plugin is disabled or uninstalled:
If an S3 server has been configured, the service is not affected.
Restart the project to automatically switch to inter-node synchronization.
The file server is abnormal.
Because S3 is essentially a KEY-VALUE database, there is no concept of path, and changes to the parent directory will recurse to all subdirectories and subfiles, so it performs poorly in functions such as backup and restore that involve frequent folder changes.