Install GPU-enabled Milvus

Prerequisites

System requirements

Operating system Supported versions
CentOS 7.5 or higher
Ubuntu LTS 18.04 or higher

Hardware requirements

Component Recommended configuration
CPU Intel CPU Haswell or higher
GPU NVIDIA Pascal or higher
RAM 8 GB or more (depends on data volume)
Hard drive SATA 3.0 SSD or higher

Milvus Docker requirements

Step 1 Confirm Docker status

Confirm that the Docker daemon is running in the background:

$ docker info

If you do not see the server listed, start the Docker daemon.

Note: On Linux, Docker needs sudo privileges. To run Docker command without sudo, create the docker group and add your user. For details, see the post-installation steps for Linux.

Step 2 Pull Milvus image

Pull the GPU-enabled image:

$ docker pull milvusdb/milvus:0.8.0-gpu-d041520-464400

Note: If the pulling speed is too slow or the pulling process constantly fails, refer to Operational FAQ for possible solutions.

Step 3 Download configuration files

# Create Milvus file
$ mkdir -p /home/$USER/milvus/conf
$ cd /home/$USER/milvus/conf
$ wget https://raw.githubusercontent.com/milvus-io/milvus/v0.8.0/core/conf/demo/server_config.yaml
$ wget https://raw.githubusercontent.com/milvus-io/milvus/v0.8.0/core/conf/demo/log_config.conf

Note: In case you encounter problems downloading configuration files using wget command, you can also create the server_config.yaml and log_config.conf files under /home/$USER/milvus/conf, then copy and paste the content from server config file and log config file.

Step 4 Start Docker container

Before starting the Docker container, you must set enable to true in gpu_resource_config section of server_config.yaml.

# Start Milvus
$ docker run -d --name milvus_gpu --gpus all \
-p 19530:19530 \
-p 19121:19121 \
-p 9091:9091 \
-v /home/$USER/milvus/db:/var/lib/milvus/db \
-v /home/$USER/milvus/conf:/var/lib/milvus/conf \
-v /home/$USER/milvus/logs:/var/lib/milvus/logs \
-v /home/$USER/milvus/wal:/var/lib/milvus/wal \
milvusdb/milvus:0.8.0-gpu-d041520-464400

The docker run options used in the above command are defined as follows:

  • -d: run container in background and print container ID
  • --name: assign a name to the container
  • --gpus: GPU devices to add to the container (‘all’ to pass all GPUs)
  • -p: publish a container’s port(s) to the host
  • -v: mounts a directory into the container

Confirm Milvus running status by the following command:

# Confirm Milvus status
$ docker ps

If Milvus server is not successfully started, you can check the error logs by the following command.

# Get id of the container running Milvus
$ docker ps -a
# Check docker logs
$ docker logs <milvus container id>

What's next

Edit
© 2019 - 2020 Milvus. All rights reserved.