Building Guest Images for OpenStack Trove

Overview

When Trove receives a command to create a database instance, it does so by launching a Nova instance based on the appropriate guest image that is stored in Glance. This document shows you the steps to build the guest images.

Note

For testing purpose, the Trove guest images of some specific databases are periodically built and published in http://tarballs.openstack.org/trove/images/ in Trove upstream CI.

Since Victoria release, Trove supports to run database service as docker container inside the guest instance, so that we don’t need to maintain multiple images for different database service. That’s the reason that you can see images for MySQL and MariaDB for Ussuri and Train releases in http://tarballs.openstack.org/trove/images/.

Additionally, if you install Trove in devstack environment, the guest image is created and registered in Glance automatically, unless it’s disabled by setting TROVE_ENABLE_IMAGE_BUILD=false in devstack local.conf file.

High Level Overview of a Trove Guest Instance

At the most basic level, a Trove Guest Instance is a Nova instance launched by Trove in response to a create command. This section describes the various components of a Trove Guest Instance.

Operating System

The officially supported operating system is Ubuntu, based on which the functional tests are running.

Docker

Since Vitoria, the database service is running as docker container inside the trove guest instance, so docker should be installed when building the guest image. This also means the trove guest instance should be able to pull docker images from the image registry(either from user port or trove management port), the related options for container images are:

[mysql]
docker_image
backup_docker_image

[postgresql]
docker_image
backup_docker_image

[mariadb]
docker_image
backup_docker_image

Trove Guest Agent

The guest agent runs inside the Nova instances that are used to run the database engines. The agent listens to the messaging bus for the topic and is responsible for actually translating and executing the commands that are sent to it by the task manager component for the particular datastore.

Trove guest agent is responsible for datastore docker container management.

Injected Configuration for the Guest Agent

When TaskManager launches the guest VM it injects config files into the VM, including:

  • /etc/trove/conf.d/guest_info.conf: Contains some information about the guest, e.g. the guest identifier, the tenant ID, etc.

  • /etc/trove/conf.d/trove-guestagent.conf: The config file for the guest agent service.

In addition to these config files, Trove supports to inject user data when launching the instance for customization on boot time, e.g. network configuration, hosts file settings, etc. The user data files are located inside the directory configured by cloudinit_location, for mysql, the file name is mysql.cloudinit

Persistent Storage, Networking

The database stores data on persistent storage on Cinder (if CONF.volume_support=True) or ephemeral storage on the Nova instance. The database service is accessible over the tenant network provided when creating the database instance.

The cloud administrator is able to config management networks(CONF.management_networks) that is invisible to the cloud tenants, but used for communication between database instance and the control plane services(e.g. the message queue).

Building Guest Images

Since Victoria release, a single trove guest image can be used for different datastores, it’s unnecessary to maintain different images for differnt datastores.

Build images using trovestack

trovestack is the recommended tooling provided by Trove community to build the guest images. Before running trovestack command:

git clone https://opendev.org/openstack/trove
cd trove/integration/scripts

The trove guest image could be created by running the following command:

$ ./trovestack build-image \
    ${guest_os} \
    ${guest_os_release} \
    ${dev_mode} \
    ${guest_username} \
    ${output_image_path}
  • Currently, only guest_os=ubuntu and guest_os_release=bionic are fully tested and supported.

  • Default input values:

    guest_os=ubuntu
    guest_os_release=bionic
    dev_mode=true
    guest_username=ubuntu
    output_image_path=$HOME/images/trove-guest-${guest_os}-${guest_os_release}-dev.qcow2
    
  • dev_mode=true is mainly for testing purpose for trove developers and it’s necessary to build the image on the trove controller host, because the host and the guest VM need to ssh into each other without password. In this mode, when the trove guest agent code is changed, the image doesn’t need to be rebuilt which is convenient for debugging. Trove guest agent will ssh into the controller node and download trove code during the service initialization.

  • If dev_mode=false, the trove code for guest agent is injected into the image at the building time.

  • Some other global variables:

    • HOST_SCP_USERNAME: Only used in dev mode, this is the user name used by guest agent to connect to the controller host, e.g. in devstack environment, it should be the stack user.

  • The image type can be easily changed by specifying a different image file extension, e.g. to build a raw image, you can specify $your-image-name.raw as the output_image_path parameter.

For example, in order to build a guest image for Ubuntu Bionic operating system in development mode:

$ ./trovestack build-image ubuntu bionic true ubuntu

Once the image build is finished, the cloud administrator needs to register the image in Glance and register a new datastore or version in Trove using trove-manage command, e.g. after building an image for MySQL 5.7.29:

$ openstack image create trove-guest-ubuntu-bionic \
  --private \
  --disk-format qcow2 \
  --container-format bare \
  --file ~/images/trove-guest-ubuntu-bionic-dev.qcow2
$ trove-manage datastore_version_update mysql 5.7.29 mysql $image_id "" 1
$ trove-manage db_load_datastore_config_parameters mysql 5.7.29 ${trove_repo_dir}/trove/templates/mysql/validation-rules.json

Note

The command trove-manage needs to run on Trove controller node. Otherwise, you can use openstack datastore version create CLI.

If you see anything error or need help for the image creation, please ask help either in #openstack-trove IRC channel or sending emails to openstack-discuss@lists.openstack.org mailing list.