Request Tracker

rt is the OCF's ticketing system, used both for internal tickets as well as for our public email lists.

This repo is designed to be an example of a simple OCF service running on Mesos. RT is an ideal service to start with because:

All state is isolated into the database, so assuming each instance can access the same DB, they don't need to coordinate in any way.
Okay to run multiple instances at once and kill them at any time (unlike other services like IRC which can only tolerate one instance, and don't cope well with that instance randomly dying).

Upgrading

Deploying new major releases of RT involves more work than just building a new Docker image and deploying that on Marathon. Here are some general instructions for updating Request Tracker:

Read the release notes to find out what you might need to change in the RT configuration. In particular, you'll want to read the README and UPGRADING-x.y files in the RT documentation.
Make any necesary configuration changes. Most configuration changes will go into the 99-ocf.pm file. You may also have to change other files as well.
Run make test to perform a smoke test. This will test to ensure that the Docker image can be built and database setup works. Just because you passed the test, though, doesn't mean it necessarily works yet.
Backup and upgrade the database. First, backup the RT database with
```
mysqldump -u ocfrt ocfrt > ocfrt.sql
```
This gives you a backup of the RT database at ocfrt.sql, in case something goes very wrong.

Next, upgrade the RT database with
```
rt-setup-database --action upgrade --dba ocfrt
```
You'll need the password for the ocfrt account, which you can find under /opt/share/docker/secrets/rt on the hypervisors.
Push your changes to GitHub and let Marathon deploy them. Test to make sure everything is working, and go back and make any changes as necessary.

dkess/rt

Request Tracker

Upgrading