Upgrade from 3.1.x

  1. Upgrade the dependencies

  2. Perform the migrations management command; in case some attribute is conflicting, remove it manually from the DB

  3. Proform the collectstatic management command

  4. Perform the set_all_layers_metadata -d management command

  5. Drop the rabbitmq image and volume and let GeoNode recreate the queues automatically

Upgrade from 2.10.x / 3.0

Upgrade the instance dependencies

Check the 1. Install the dependencies and 2. GeoNode Installation sections in order to upgrade your Python environment.

Also, make sure the code is Python 3.8 compatible and that you switched and aligned the source code and the requirements.txt to the 3.x branch.

This must be done manually and with particular attention.

workon geonode3
cd /<full_path_to_geonode>

pip install pip --upgrade
pip install -r requirements.txt --upgrade
pip install -e . --upgrade
pip install pygdal=="`gdal-config --version`.*"

./manage.sh collectstatic --noinput

Prepare the Database and Migrate to the new Schema

Fix the tables in order to migrate to the new schema

./manage.sh dbshell

ALTER TABLE base_resourcebase ADD COLUMN doi_bkp varchar;
UPDATE base_resourcebase SET doi_bkp = doi;
ALTER TABLE base_resourcebase DROP COLUMN doi;

CREATE TABLE base_backup(name varchar);

CREATE TABLE base_usergeolimit_bkp ( like base_usergeolimit including all);
CREATE TABLE base_groupgeolimit_bkp ( like base_usergeolimit including all);
CREATE TABLE base_resourcebase_users_geolimits_bkp ( like base_usergeolimit including all);
CREATE TABLE base_resourcebase_groups_geolimits_bkp ( like base_usergeolimit including all);

DROP TABLE IF EXISTS base_configuration CASCADE;
DROP TABLE IF EXISTS base_usergeolimit CASCADE;
DROP TABLE IF EXISTS base_groupgeolimit CASCADE;
DROP TABLE IF EXISTS base_resourcebase_users_geolimits CASCADE;
DROP TABLE IF EXISTS base_resourcebase_groups_geolimits CASCADE;

\q

Migrate to the new schema

./manage.sh makemigrations
./manage.sh migrate

Restore the old contents

./manage.sh dbshell

UPDATE base_resourcebase SET doi = doi_bkp;
ALTER TABLE base_resourcebase DROP COLUMN doi_bkp;

INSERT INTO base_usergeolimit (SELECT * FROM base_usergeolimit_bkp);
INSERT INTO base_groupgeolimit (SELECT * FROM base_groupgeolimit_bkp);
INSERT INTO base_resourcebase_users_geolimits (SELECT * FROM base_resourcebase_users_geolimits_bkp);
INSERT INTO base_resourcebase_groups_geolimits (SELECT * FROM base_resourcebase_groups_geolimits_bkp);

DROP TABLE IF EXISTS base_usergeolimit_bkp CASCADE;
DROP TABLE IF EXISTS base_groupgeolimit_bkp CASCADE;
DROP TABLE IF EXISTS base_resourcebase_users_geolimits_bkp CASCADE;
DROP TABLE IF EXISTS base_resourcebase_groups_geolimits_bkp CASCADE;

\q

Upgrade from 2.4.x

These are the notes of a migration from 2.4.x to 2.10.1. These notes could possibly work also when migrating from 2.6.x, 2.7.x, 2.8.x but are not tested in that scenarios. You should run this procedure on your local machine and once you successfully migrated the database move the backup to your GeoNode 2.10.1 production instance.

PostgreSQL

Create a role and a database for Django GeoNode 2.4:

create role user with superuser login with password '***';
create database gn_24 with owner user;
\c gn_24
create extension postgis;

Restore backup from your production backup:

psql gn_24 < gn_24.sql

Run GeoNode migrations

Activate your GeoNode virtualenv and set the env vars:

. env/bin/Activate
export vars_210

Here are the variables to export - update them to your environment settings:

export DATABASE_URL=postgis://user:***@localhost:5432/dbname
export DEFAULT_BACKEND_DATASTORE=data
export GEODATABASE_URL=postgis://user:***@localhost:5432/geonode_data
export ALLOWED_HOSTS="['localhost', '192.168.100.10']"
export STATIC_ROOT=~/www/geonode/static/
export GEOSERVER_LOCATION=http://localhost:8080/geoserver/
export GEOSERVER_PUBLIC_LOCATION=http://localhost:8080/geoserver/
export GEOSERVER_ADMIN_PASSWORD=geoserver
export SESSION_EXPIRED_CONTROL_ENABLED=False

Downgrade psycopg2:

pip install psycopg2==2.7.7

Apply migrations and apply basic fixtures:

cd wfp-geonode
./manage.py migrate --fake-initial
paver sync

Regenerate from scratch the upload application tables in the database:

delete from django_migrations where app = 'upload';
drop table upload_upload cascade;
drop table upload_uploadfile;

Regenerate upload tables with migrate:

./manage.py migrate upload

Upgrade psycopg2:

pip install -r geonode/requirements.txt

Create superuser

To create a superuser you should drop the following constraints (they can be re-enabled if needed):

alter table people_profile alter column last_login drop not null;

./manage createsuperuser

Fixes on database

For some reason some resources were unpublished:

UPDATE base_resourcebase SET is_published = true;

Remove a foreign key from account_account which is not used anymore (GeoNode dev team: maybe even better let’s remove all of the account tables, I think they are stale now):

ALTER TABLE account_account DROP CONSTRAINT user_id_refs_id_726cb6b4;
ALTER TABLE account_signupcode DROP CONSTRAINT "inviter_id_refs_id_49a7c0d9";

Fix the remote service layers by running this script:

python migration/fixes_remote_layers.py