We have been using postgres from the very start of our company – Directi. The talk will give an overview of how our experience has been so far, alongwith the configuration of postgres and auxilliary components that we have in place. The talk will cover the following topics:
* >10 years of Postgres in production
. The topic will cover our experience with postgres in production, size of database and it’s growth over a period of time, the type of workload and QPS our postgres r/w DB caters to.
* Configuration of Postgres Cluster
The topic will cover the configuration of postgres cluster in production. We’re using physical streaming replication provided by postgres to have one master and multiple slaves in the cluster. Different application access the RO or RW DB as per the requirements.
* Database Failover Process
The topic will cover the automated process we follow for performing Database Failover. We’re not using any tool for performing failover, and we have rather built an automated process in-house, since the database is the most critical component of our application in the system architecture.
* Performance tuning parameters
This topic will contain a brief overview of the performance tuning parameters that we have configured specific to our production workload alongwith the rationale behind the same.
* Backups, Recovery, RTO, RPO, Avoiding Database corruption
This topic will cover other misc configuration we have around postgres backups and recovery. We have a specific process that runs everyday to ensure the database is not corrupted at any point of time.
Demos: (If time permits)
* Quick demo on how to setup streaming replication with postgres
* Quick demo of our Database Failover process