On 15/04/13 09:49, Anand Kumria wrote:
Hi,
I've been evaluating Bloodhound, and like what I see.
And now I want to roll it out to multiple - independent - users.
Is it better to have a single installation (i.e. one virtualenv), or one
(virtualenv) per user?
I presume that each user should also get their own database as well.
Hi Anand,
I might be missing something but I am not sure we answered all your
questions.
I would have thought it would be most convenient for you to go with a
single virtualenv and multiple bloodhound environments. As the clients
are properly independent, the advantages of hosting in a single database
are probably lost so you probably do not want multiproduct to help with
that aspect.
One of the bigger advantages to running a different virtualenv per
client is that, when you want to upgrade, it naturally gives you the
ability to update each of your clients independently. However, that
flexibility is also possible to achieve with running on one virtualenv
for the current version and when it comes to upgrade time, install the
new version in a separate virtualenv and migrate each user's bloodhound
environments to it one at a time. (Obviously making sure you are backed
up and also testing that the migration will work on a staging server
would be advisable.) You can always choose to adjust the number of
virtualenvs later if you really find a need to.
As it is possible to specify both the location of a directory in which
to put your bloodhound environments (--environments-directory) and the
bloodhound environment name (--project) in the bloodhound_setup.py
script, you may want to make use of those so that all bloodhound
environments are kept in an obvious independent location.
Hope that makes sense.
Cheers,
Gary
