replisim

Data Replicating Appliance — generates realistic, continuous database and filesystem loads for Disaster Recovery (DR) replication testing.

It simulates the kind of production-like data churn that real DR testing requires, running as long-lived background processes (systemd services).

---

What it does

Database load (db_changer.py)

Continuously inserts batches of fake records into a MySQL users table using the Faker library. Enforces a row cap by deleting the oldest records when the limit is hit, maintaining a rolling window of data.

Filesystem load (file_changer.sh)

Continuously writes random binary files (via /dev/urandom) to a target directory. When the directory exceeds a configurable size limit, it deletes the oldest files to stay under the threshold.

---

Components

File	Purpose
db_changer.py	Main DB load generator
db_changer.ini	Config: DB credentials, batch size, field definitions
db_wipe.py	Utility to drop all tables and reset the DB
file_changer.sh	Filesystem load generator
file_changer.ini	Config: max dir size, chunk size, sleep interval, target path
wiper.sh	Stops services, deletes logs/data, wipes DB, clears history

---

Configuration

All tunable parameters live in .ini files — no code changes needed to adjust behavior.

db_changer.ini defaults:

• max_rows: 100,000
• records_per_batch: 5,000
• sleep_time: 30 seconds
• Fields: name, age, email, city, created_at, is_active, description, random_number

file_changer.ini defaults:

• MAX_SIZE: 10G
• CHUNK_SIZE: 500M
• SLEEP_TIME: 30 seconds
• DIR: /opt/replisim/data

---

Design notes

• Schema-flexible: db_changer.py reads field definitions from config and dynamically creates or migrates the MySQL table, including adding new columns to existing tables.
• Bounded operation: Both tools implement rolling windows — the DB stays under max_rows and the directory stays under MAX_SIZE — so they run indefinitely without exhausting disk or database space.
• Faker-backed: DB records use realistic fake data (names, emails, cities, dates, etc.) to more closely mimic real replication traffic.

---

Tech stack

• Python 3 + mysql.connector + faker
• Bash + dd + du
• MySQL
• systemd