We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
DOs and DON'Ts of managing numerous very large databases at CERN by Andrzej Nowicki
Management nightmares: Learn how CERN automates and optimizes its large Oracle, MySQL, and Postgres databases, including use cases for physical standby replication, user management, and monitoring/alerting systems.
- Andrzej Nowicki speaks about managing large databases at CERN, describing it as a “nightmare” due to the complexity and scale.
- The team at CERN uses automation to manage databases, with a focus on Oracle and MySQL/Postgres.
- They have implemented a system for monitoring and alerting on database issues.
- The team uses physical standby databases for replication and cloning.
- They also use file tablespaces for storage, which can grow automatically.
- In terms of user management, CERN uses a self-service portal to allow users to manage their own database accounts.
- The team also uses Rundeck for database automation and management.
- Despite these efforts, the team still faces challenges in managing their large databases, including issues with circular dependencies and the need for backup and disaster recovery.
- CERN is working on improving their monitoring and alerting systems to better detect and respond to issues.
- The team is also exploring new technologies, such as Elastic, to help manage their large databases.
- Andrzej highlights the importance of having a centralized logging system to track database activity.
- He also emphasizes the need to protect critical infrastructure from automatic deletion.
- The team is actively working on providing a solution for user requests to manage their own databases.
- Andrzej mentions that they use a combination of snapshots and backups to ensure data recovery.
- He also provides information on the growth of CERN’s databases, from 1 exabyte to 1.5 petabytes.
- CERN’s databases include Oracle, MySQL, Postgres, and Influx, and are used for various purposes, including data storage and analysis.