Banco de dados em Containers Fernando Ike PGConf.Brasil 2018 

Cache/Buffer 

Cache An auxiliary memory from which high-speed retrieval is possible. https://en.oxforddictionaries.com/definition/cache 

Buffer A temporary memory area in which data is stored while it is being processed or transferred, especially one used while streaming video or downloading audio. https://en.oxforddictionaries.com/definition/buffer 

https://momjian.us/main/writings/pgsql/hw_performance/ 

No content

Database cache Filesystem cache Disk cache Store data (disk) 

What’s is Virtual Machine? “...an efficient, isolated duplicate of a real computer machine...” Gerald Popek and Robert Goldberg 

VM Infrastructure Hypervisor App A Artifact Bin/Libs App B Bin/Libs App C Bin/Libs Guest OS Guest OS Guest OS 

What’s container? Container is a type of virtualization, more common call “operating-system-level virtualization” 

Container Infrastructure Hypervisor Docker App A Artifact Bin/Libs App B Bin/Libs App C Bin/Libs 

Ephemera Ephemera (singular: ephemeron) are any transitory written or printed matter not meant to be retained or preserved 

Stateless application Applications don’t need to store anything in a persistent disk. Data application are storage in other part of architecture system 

Stateful application Application that need to store “locally” data and need to recover that easy after an “interruption” 

Cap Theorem 

Union filesystem It allows files and directories of separate file systems, known as branches, to be transparently overlaid, forming a single coherent file system. 

Container layers 

Main filesytem for internal container data AUFS Overlayfs Overlayfs2 Device-Mapper Bind/Mount 

Databases operation Plan what type of workload to use Maintain, organize and “improve” dataset together application Install security fix releases Archive “old” data (sanitize) Upgrade Database version Tuning OS, Database and Application 

What’s means in OS handle... sysctl /proc /sys limits overcommit 

What’s means in the PostgreSQL Shared Buffers WAL Max_connections Effective_cache_size Checkpoint_segments Random_page_cost 

Container Volumes They are a mechanism for persisting data generated and/or used by container 

Why to use PostgreSQL within containers Lower footprint to start Easy to change postgresql.conf and use git to have history change (tracking) Less surface attack in the “application Can “restrict” Linux Capabilities within container “Small” database by microservice Use always a Container Storage 

Tips One database per type of workload Take care to restrict network, consider restrict database access only own application Push database log to Monitoring system/service Clusters and data replication are more complex Database use intensive CPU and IO, think if there is another service with the same workload type Database is prepare to support two version of application (Deploy strategies without downtime) 

version: "3" services: db: image: postgres environment: - POSTGRES_USER=postgres - POSTGRES_PASSWORD=postgress - POSTGRES_DB=postgres ports: - "5433:5432" volumes: - pgdata:/var/lib/postgresql/data networks: - postgres volumes: pgdata: networks: postgres: 

Demo 

Demo code https://gist.github.com/fike/158204255cfbc368f36fad66ccd999a7 

● Bruce Momjian - Performance tuning ● Docker Volumes ● Kubernetes Volumes ● Oxford Dictionaries ● Formal requirements for virtualizable third generation architectures - Gerald J. Popek and Robert P. Goldberg ● CAP Theorem and Distributed Database Management Systems References 

// http://www.10deploys.com // [email protected] // https://twitter.com/fernandoike // http://www.fernandoike.com // https://www.linkedin.com/in/fernandoike/ Contacts