Using System Catalogs to Investigate a PostgreSQL Server

PostgreSQL is the world’s most advanced open source database. Its stability, functionality, and extensibility make it a primary choice for an RDBMS solution.

In this hands-on lab scenario you are the DBA for Awesome Company. You have built a PostgreSQL database backend to facilitate the development of a new web application. Now that it is being utilized by the development group, you need to gather some diagnostic information about the instance.

Performing the tasks of this lab will help you become familiar with utilizing the system catalogs to gather information about your PostgreSQL installation. This includes finding and evaluating sessions that are generating load, listing the size of objects, and querying information about column types.

Learning Objectives

Successfully complete this lab by achieving the following learning objectives:

Prepare the Environment

Restore a backup:
- Change to the postgres user:
```
sudo su - postgres
```
- Create the database to be restored:
```
createdb acweb
```
- Download the backup from GitHub:
```
wget https://github.com/linuxacademy/content-postgresql-deepdive/raw/master/acweb/acweb.tar
```
- Use pg_restore to restore the backup:
```
pg_restore --dbname=acweb --verbose /var/lib/pgsql/acweb.tar
```
- Launch psql and verify the data is present:
```
psql
```
- List the databases with l.
- Connect to acweb:
```
c acweb
```
- Get a count for the payment table:
```
SELECT COUNT(*) FROM sales.payment;
```
- Exit psql:
```
q
```
Generate load against the database:
- Download a script to generate load against the database:
```
wget https://raw.githubusercontent.com/linuxacademy/content-postgresql-deepdive/master/acweb/loadgen.sh
```
- Modify the file to be executable:
```
chmod +x loadgen.sh
```
- Execute the script in the background:
```
./loadgen.sh > /dev/null 2>&1 &
```

Evaluate Load-Generating Sessions

Launch psql and turn on expanded display:
```
psql
x
```

Execute the query below against the pg_stat_activity system catalog:

SELECT pid, datname, usename, application_name, client_addr, xact_start, wait_event_type, wait_event, state, query
FROM pg_stat_activity WHERE application_name = 'psql';

Note in particular the wait information. You may have to run it a few times, but you should begin to see evidence of both write and read I/O.

List the Size of the Database and Tables

Connect to the acweb database:
```
c acweb
```

Execute the following query to acquire the database size:

SELECT pg_size_pretty(pg_database_size(current_database()));

Execute the following query to acquire the size of all tables in the sales schema:

SELECT table_schema, table_name, pg_size_pretty(pg_relation_size('"'||table_schema||'"."'||table_name||'"')) AS size
FROM information_schema.tables
WHERE table_schema = 'sales'
ORDER BY size DESC;

Find All of the VARCHAR Columns

Execute the following query to list all columns of the type VARCHAR in the sales schema:

SELECT columns.attname as name,
data_types.typname as type,
class.relname as table,
tables.schemaname as schema
FROM pg_attribute columns
INNER JOIN pg_class class ON columns.attrelid = class.oid
INNER JOIN pg_tables tables on class.relname = tables.tablename
INNER JOIN pg_type data_types 
ON columns.atttypid = data_types.oid
WHERE tables.schemaname = 'sales'
AND data_types.typname = 'varchar';

Note how we can join multiple system catalogs together in order to obtain more useful information.

What are Hands-on Labs

Hands-on Labs are real environments created by industry experts to help you learn. These environments help you gain knowledge and experience, practice without compromising your system, test without risk, destroy without fear, and let you learn from your mistakes. Hands-on Labs: practice your skills before delivering in the real world.

Using System Catalogs to Investigate a PostgreSQL Server

About this Hands-on Lab

Learning Objectives

Additional Resources

What are Hands-on Labs

Newsletter