These best practices warrant consideration, but are not expected to apply to every business or in every situation. Recommendations here are not mandatory.
We recommend that you:
Understand business requirements before deployment.
Adapt deployment practices to align to business requirements.
Consider to what extent deployed systems should integrate with existing business systems and practices.
Servers should be validated before exposure to production workloads and production data.
It may be appropriate to prevent access to an unconfigured server until configured and validated. Load balancer configuration, firewall rules, or database server configuration are often used to prevent unintended traffic to new servers.
Details assessed during Pre-Production validation can include:
Server capacity, including performance of disk systems
Server security configuration and hardening
Tuning for initial data load vs production workloads
Alignment to business requirements
Change Management, Automation, Orchestration¶
Server configuration changes should be done through change management. Accurate records of the time of change and reason for change can enable faster issue diagnosis.
Automation can enable repeatability of change deployment, and can aid Pre-Production and Post-Production testing.
Automation or orchestration can enable repeatability of server deployment, including system provisioning.
Database servers exist to run database server software. Database servers should not also run web server software, act as a file server, or run other workloads.
Understand workload and data isolation requirements before server deployment. Isolation requirements are often defined through business requirements, including:
Data and application security requirements that trigger isolation of one workload from other workloads.
Separation of Development and Testing environments from Production environments.
Understand control requirements before server deployment, including records of control implementation needed to support audits.
Control requirements typically follow data:
Both production and non-production systems may require production-level controls based on the presence of data subject to control.
Database servers, backup servers, and other systems may require production-level controls based on the presence of data subject to control.
MariaDB Product Notifications allow you to keep aware of new releases, including security fixes. Customers can manage MariaDB Product Notifications through the MariaDB Customer Portal.