We need to separate the OpenQuake data per user/group

[Muharem Hrnjadovic]

This involves all kinds of data: RabbitMQ/celery artefacts, files on disk etc.

When we introduce user and group management ("We need to add authentication, permissions and quotas", https://github.com/gem/openquake/issues/272) we will have to keep the OpenQuake data of these users/groups separated for security reasons.

For celery separate RabbitMQ vhosts should probably do the trick.

Adding this capability may also help with issue https://github.com/gem/openquake/issues/273 ("We need to be able to deploy non-master branches")

[Lars Butler (lars-butler)]

No longer relevant.