This repository holds all the scripts that are run on the nightly performance environment. The environment consists out of:
- 3 Large (8GB RAM) centOS machines running Cassandra 1.1.5
- 2 Small (1GB RAM) SmartOS machines running the app servers
- 1 small (1GB RAM) SmartOS machine running the load balancer (nginx)
Each night the following task will be run:
- Stop all the cassandra nodes and restore a snapshot. This snapshot holds 40000 users, 80000 groups and 200000 content items.
- Stop the app nodes and redeploy latest master.
- Generate 10 batches of which each batch holds 1000 users, 2000 groups and 5000 content items
- Load those batches in.
- Package the new data and generate Tsung tests (See node-oae-tsung)
- Run the tsung tests.
- Publish all the collected data to Circonus and in the www directory of the driver machine.
- Run a little script that provides some raw data metrics.