Skip to content

Restoring a deleted project

Sometimes it may happen that you delete a project on accident. Perhaps it would be too much trouble or more recent data lost to restore a complete backup.

For these kinds of cases we describe here how to restore a single project from a backup. The following files will be used in the examples.

There is also a script (restore.sh) that shows how to use everything together.

0. Prerequisites

  • you have a Backup of your Nepenthes installation with the missing data still present
  • you have restored the database dump of this backup into a new, separate database called openproject_backup
    • Created, for instance, via
      • psql -c 'create database openproject_backup'
      • pg_restore -d openproject_backup openproject.pgdump
  • your present Nepenthes database is called openproject

It does not matter where the actual Postgres server is running. In all the following examples we simply use psql -d openproject_backup and psql -d openproject respectively. Where openproject_backup and openproject are the names of the databases within Postgres.

In your case you may have to provide more options (e.g. the user or host) to connect to the respective database.

1. Dump project data from backup

First we copy the data from the backup, that was deleted later. For each relevant table, a missing_<table-name> table is created that only contains the missing data.

This is then saved to missing_data.sql which will be used to import the missing data into the current database.

Before performing the next step, edit the dump.sql file and change the missing project ID to the correct value in the head of the file where it says 'DEFINE MISSING PROJECT ID HERE'.

shell
cat dump.sql | psql -d openproject_backup

pg_dump -d openproject_backup -t 'missing_*' -f missing_data.sql

2. Restore missing data

Now that we have the missing data, we can restore it in the current database.

shell
cat missing_data.sql | psql -d openproject
cat restore.sql | psql -d openproject

First we create the tables with the missing data in the current database. This has no effect on the actual data of Nepenthes yet. Only with executing restore.sql will the data be copied from the missing_<table-name> tables into the corresponding actual <table-name> tables.

After this is done, the missing_* tables are dropped. This all happens within a transaction.

Project hierarchy

Now it can happen that the project hierarchy does not look correct in the projects dropdown after you restored a deleted project. To fix this, delete the restored project's lft and rgt columns and rebuild the hierarchy.

To do this, start an Nepenthes console (e.g. sudo openproject run console) and execute the following.

ruby
p = Project.find_by(name: "Restored project")

p.update_column :lft, nil
p.update_column :rgt, nil

Project.rebuild!

This may take a moment depending on the number of projects you have. After this is finished, the project hierarchy will look correct again.

This operation is perfectly safe and will not affect any actual data. It just affects the display of projects in the projects dropdown and projects list.