This application enables django powered websites to have multiple tenants via PostgreSQL schemas. A vital feature for every Software-as-a-Service (SaaS) website.
Read the full documentation here: django-tenants.readthedocs.org
Django provides currently no simple way to support multiple tenants using the same project instance, even when only the data is different. Because we don’t want you running many copies of your project, you’ll be able to have:
- Multiple customers running on the same instance
- Shared and Tenant-Specific data
- Tenant View-Routing
A schema can be seen as a directory in an operating system, each directory (schema) with its own set of files (tables and objects). This allows the same table name and objects to be used in different schemas without conflict. For an accurate description on schemas, see PostgreSQL’s official documentation on schemas.
There are typically three solutions for solving the multitenancy problem.
- Isolated Approach: Separate Databases. Each tenant has its own database.
- Semi Isolated Approach: Shared Database, Separate Schemas. One database for all tenants, but one schema per tenant.
- Shared Approach: Shared Database, Shared Schema. All tenants share the same database and schema. There is a main tenant-table, where all other tables have a foreign key pointing to.
This application implements the second approach, which in our opinion, represents the ideal compromise between simplicity and performance.
- Simplicity: barely make any changes to your current code to support multitenancy. Plus, you only manage one database.
- Performance: make use of shared connections, buffers and memory.
Each solution has its up and down sides. For a more in-depth discussion, see Microsoft’s excellent article on Multi-Tenant Data Architecture.
Tenants are identified via their host name (i.e tenant.domain.com). This
information is stored on a table on the public
schema. Whenever a
request is made, the host name is used to match a tenant in the
database. If there’s a match, the search path is updated to use this
tenant’s schema. So from now on all queries will take place at the
tenant’s schema. For example, suppose you have a tenant customer
at
http://customer.example.com. Any request incoming at
customer.example.com
will automatically use customer
’s schema
and make the tenant available at the request. If no tenant is found, a
404 error is raised. This also means you should have a tenant for your
main domain, typically using the public
schema. For more information
please read the setup section.
Each tenant has its data on a specific schema. Use a single project instance to serve as many as you want.
Tenant-specific apps do not share their data between tenants, but you can also have shared apps where the information is always available and shared between all.
You can have different views for http://customer.example.com/
and
http://example.com/
, even though Django only uses the string after
the host name to identify which view to serve.
Everyone loves magic! You’ll be able to have all this barely having to change your code!
This is just a short setup guide. It is strongly recommended that you read the complete version at django-tenants.readthedocs.org.
Your DATABASE_ENGINE
setting needs to be changed to
DATABASES = {
'default': {
'ENGINE': 'django_tenants.postgresql_backend',
# ..
}
}
Add the middleware django_tenants.middleware.main.TenantMainMiddleware
to the
top of MIDDLEWARE
, so that each request can be set to use
the correct schema.
MIDDLEWARE = (
'django_tenants.middleware.main.TenantMainMiddleware',
#...
)
Add django_tenants.routers.TenantSyncRouter
to your DATABASE_ROUTERS
setting, so that the correct apps can be synced depending on what's
being synced (shared or tenant).
DATABASE_ROUTERS = (
'django_tenants.routers.TenantSyncRouter',
)
Add django_tenants
to your INSTALLED_APPS
.
from django.db import models
from django_tenants.models import TenantMixin, DomainMixin
class Client(TenantMixin):
name = models.CharField(max_length=100)
paid_until = models.DateField()
on_trial = models.BooleanField()
created_on = models.DateField(auto_now_add=True)
class Domain(DomainMixin):
pass
Define on settings.py
which model is your tenant model. Assuming you
created Client
inside an app named customers
, your
TENANT_MODEL
should look like this:
TENANT_MODEL = "customers.Client" # app.Model
TENANT_DOMAIN_MODEL = "customers.Domain" # app.Model
Now run migrate_schemas
. This will sync your apps to the public
schema.
python manage.py migrate_schemas --shared
Create your tenants just like a normal django model. Calling save
will automatically create and sync the schema.
from customers.models import Client, Domain
# create your public tenant
tenant = Client(schema_name='tenant1',
name='My First Tenant',
paid_until='2014-12-05',
on_trial=True)
tenant.save()
# Add one or more domains for the tenant
domain = Domain()
domain.domain = 'tenant.my-domain.com'
domain.tenant = tenant
domain.is_primary = True
domain.save()
Any request made to tenant.my-domain.com
will now automatically set
your PostgreSQL’s search_path
to tenant1
and public
, making
shared apps available too. This means that any call to the methods
filter
, get
, save
, delete
or any other function
involving a database connection will now be done at the tenant’s schema,
so you shouldn’t need to change anything at your views.
You’re all set, but we have left key details outside of this short tutorial, such as creating the public tenant and configuring shared and tenant specific apps. Complete instructions can be found at django-tenants.readthedocs.org.
django-tenants comes with an example project please see
I would like to thank two of the original authors of this project.
- Bernardo Pires under the name django-tenant-schemas.
- Vlada Macek under the name of django-schemata.
- Django 2 if you want to use Django 1.11 or lower please use version 1 of django-tenants
- PostgreSQL
If you want to run tests, you can either run run_tests.sh
(which requires access to
a PostgreSQL instance, location of which you can customize using the DATABASE_HOST
env variable) or use docker-compose like this:
## Start Docker service
# start docker # with Upstart
# systemctl start docker # with systemd
## Install docker-compose (you might want to do this in Python virtualenv)
# pip install docker-compose
## In main directory of this repo do:
docker-compose run --rm django-tenants-test # runs django-tenants tests.
# dockerized PostgreSQL service is started implicitly
(note that upon first run the Dockerfile
will be built).
An online video tutorial is available on youtube.
If this project helped you reduce development time, you can give me cake :)