Deploy a time-series database (Optional)

If you need OceanBase Migration Service (OMS) to collect and display historical monitoring data, deploy a time-series database. This topic describes how to deploy a time-series database and the limits.

Limits

• A time-series database can be deployed in OMS V2.1.0 and later. Currently, InfluxDB 1.8 is deployed as a time-series database.

• You need to use port 8086 to deploy a time-series database.

• A time series database supports only one-node deployment and does not support high availability. You need to deploy only one TiDB. The time-series database supports only the single-node deployment mode and does not provide high availability. You can deploy the time-series database and OMS on the same host. When the service fails, monitoring data becomes unavailable.

Procedure

1. Log on to the host where the time-series database is to be deployed.

2. Load the image of InfluxDB 1.8.

   Click here to download the InfluxDB image, and upload it to the host where the time-series database is to be deployed.

   sudo docker load -i {influxdb_image}
   

3. Start the Docker container and ensure that data can be written to the mounted data disk.

   sudo docker run  -dit -p 8086:8086 -p 14444:14444 \
   -v {Mounted data disk}:/var/lib/influxdb \
   --env INFLUXDB_BIND_ADDRESS=127.0.0.1:14444 \
   --env INFLUXDB_HTTP_AUTH_ENABLED=true \
   --env INFLUXDB_HTTP_PING_AUTH_ENABLED=true \
   --name=oms-influxdb \
   influxdb:1.8
   

4. To enable authentication for the InfluxDB service, create an account and set the password.

   1. Access the InfluxDB container.

      docker exec -it ${INFLUXDB_NAME} bash
      

   2. Access the InfluxDB console.

      cd /usr/bin
      ./influx
      

   3. Create an account, set the password, and grant the admin privileges to the account.

      Notice:

      • You must first create the admin user in InfluxDB before you enable authentication.

      • Save the account and password for configuration of the config.yaml file.

      create user "${USER_NAME}" with password '${PASSWORD}' with all privileges
      

   4. View users.

      show users;
      

5. Exit the container where the time-series database resides.