[GeoNode-users] R: migration suggestions

[Eugenio Trumpy]

Dear Toni,
in blue my comment to your answer:

I would say this depends on your maps. As you can upload several layers in one shot further use importlayers
management command to import for example all shapes of a folder I tend to say: You might be faster by just rebuilding
your data.

Some map is simple while others include many layer.
I understand that you are warmly suggesting to start from scratch.

I've never tried multisites. But from what I know it's currently not maintained.
Well I think I would fire up several docker stacks and bind them to specific domains.

This is the biggest issue to me! I never wanted to hear this. I spent time some years ago to set up the multisites for my projects. I understand that this is a non recommended solution for me now. You are suggesting me to use several docker stack, one for each project. Unfortunately, I wouldn't know from where to start, and it would require probably to much time to me to start to use docker. Any other suggestion???

The database is actually build by Django`s model definition. For example have a look
at the monitoring model:
https://github.com/GeoNode/geonode/blob/master/geonode/monitoring/models.py#L125-L137

This did not exist in 2.4 means you should not find tables for monitoring_ in a 2.4 database.
In a 2.10 database there should exist monitoring_ tables and most likely one called something
with _service having the fields as marked by above link.

Long story short. The models should allow you to spot check your migration.

As I was supposing the command didn't affect my db, I have the same table I had in the version 2.4 and consiquently
I had neither the monitoring_ tables nor _service.
Again I think there is no connection between my local_settings.py and my db so that the command 'migrate' doesn't produce any change.

Sorry for being unclear. I just confirmed your guess. As settings.py and local_settings.py define
database connection it could lead to strange behaviour when using the wrong file.

Ok, I wonder if there a way to activate local_settings.py because it doesn't affect the settings.py. If I want to change the configuration, as I did for ALLOWED_HOSTS I have to modify the settings.py. If I'm not wrong the rule was to not modify settings.py or at least at mininum and use local_settings.py for any customisation.

This depends on which user runs the apache process (I think you are using apache right)?
My apache2 is owned by www-data user

More likely there is some confusing with your used settings files.
I followed the https://docs.geonode.org/en/master/install/core/index.html when it recommends to copy paste
local_settings.py.geoserver.sample.

In case you have followed the installation guide on docs.geonode.org<http://docs.geonode.org> you would use nginx
and uwsgi. Uwsgi set s all configuration through environment variables. I would strongly suggest
to follow this setup guide ;)

I know this is the best solution, not for my server configuration at the moment. I have other application running
with Apache2. Not a walk in the part to set a second webserver (NGINX).

So, if I don't solve the issue on the local_settings.py / settings.pf files in order to migrate the db and access with the my
old administrator credential, the best solution is start from scretch. This is clearer now.
However, if I don't find a solution for the multisites, or a way to manage different projects, I need to find a way to maintain everything
at the running version and assess to use geonode new version in a dedicated server for only new project only in the future.

I'm a little bit sad, but I don't see many other solutions...

Thanks again for your support,
Eugenio


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/geonode-users/attachments/20200322/a89365b7/attachment.html>