peeringdb-peeringdb/docs/deploy.md

PeeringDB Deploy

PeeringDB currently only runs on python2.7 - all instructions below are to be taken with that in mind.

Getting Started

This document uses the following variables

export PDB_REPO=git@github.com:peeringdb/peeringdb.git

Install facsimile

pip install facsimile jinja2

Clone peeringdb

git clone $PDB_REPO

Developer instance deploymnet

Create ~/srv

mkdir ~/srv

Copy config/facsimile/dev.yaml to config/facsimile/mydev.yaml and make changes where needed.

cp config/facsimile/dev.yaml config/facsimile/mydev.yaml
vi config/facsimile/mydev.yaml

Use the facs command to deploy a virtualenv and server files for your dev instance

facs peeringdb mydev --src-dir=.
facs venv mydev --src-dir=.

Files will be deployed to ~/srv/dev.peeringdb.com (unless you changed the location in your config)

Setup passwords

Once you have run facs for the first time it will have created a .facsimile directory

You will want to open .facimsile/state/{env}/state.yaml and set the correct passwords for everything and then do another deploy to make sure the correct passwords are deployed.

instances:
  inmap: {}
  uiidmap: {}
passwd:
  deskpro: xxx # deskpro api key
  djangokey: xxx # django secret
  google_geoloc_api: xxx # google geolocation api key
  peeringdb: xxx # database password
  recaptcha: xxx # recaptcha secret

Create api-cache dir

mkdir ~/srv/dev.peeringdb.com/etc/api-cache

Symlink for convenience

In order to be able to run the manage.py command out of the pdb repository you can symlink the peeringdb_com directory from deploy location

In the peeringdb repo root:

ln -s ~/srv/dev.peeringdb.com/peeringdb/peeringdb_com peeringdb_com

Setup database

During deploy facsimile will have created a sql file at .facsimile/tmp/RELEASE/mydev/peeringdb/init.sql - load it into mysql.

mysql -u root -p < .facsimile/tmp/RELEASE/mydev/peeringdb/init.sql

Migrate database - empty, from scratch

. ~/srv/dev.peeringdb.com/venv/bin/activate
./manage.py migrate
./manage.py createcachetable
./manage.py loaddata fixtures/initial_data.json

Running the dev instance

./manage.py runserver

Populating with data

There are currently 2 ways to get some data into your developer instance

1. Sync from production peeringdb - slow, but accurate data
2. Generate test data - fast, but marginally useful test data

1) Sync from Production

You can populate your data from peeringdb.com using

./manage.py pdb_load_data

However be prepared for this to take 15-20 minutes as it will not only sync the entities, but also set up usergroups for each organization and so forth.

This should only be used to populate initial data. Once you have started adding / updating objects and your data diverges from production data, it is no longer useful to call this command.

Special Note: this will only sync data visible to everyone, any fields or rows hidden behind authentication will be missed.

2) Generate test data

Alternatively a faster way to get data into your instance is to generate a set of test data.

The data generated has no relation to real world data, but should be good enough for testing purposes

./manage.py pdb_generate_test_data --commit

Admin

The admin area can be accessed from the /cp endpoint, you will need to have a superuser to access it

./manage.py createsuperuser

Hangups

Authentication not working

This is usually caused by misconfigured session settings

In peeringdb_com/settings.d/01-local.conf

• Check that SESSION_COOKIE_DOMAIN is set to the apropriate domain
• Check that SESSION_COOKIE_SECURE is False if youre not serving over https

Versioning

Everything is versioned for deploy, using facsimile.

# to update dev versions
version_bump_dev

# to update release versions
version_merge_release

Tests

pip install -r requirements-test.txt
py.test -v -rxs --cov-report term-missing --cov=peeringdb_server/ --capture=sys tests/