Connecting to Postgres

When a new user connects to the database, postgres will fork a new process.

• After this point the client and the new process communicate directly, without intervention from the original process.
• This shows the dependency on processes rather than threads, giving us more stability at the cost of higher connection start-up costs.
• each connection into Postgres is going to consume about 10 MB of memory from your database
• These start-up costs can be overcome with pooling

In postgres, each client connecting to the database has its own process

• connection can be verified by running select true as "Connection test";

Connection String - the database url used to connect to the database (postgres:///...)

• if we specify the connection string as postgres:///mydb, then we are implicitly passing our computer's username and no password, which would be equivalent to postgres:///kyletycholiz@localhost:5432/mydb
  • spec: are these default variables determined by env variables like PGHOST?

---

Children

1. Connection Pool