I upgraded Discourse to latest stable version with complete rebuild yesterday. I’ve got this strange error this morning: [2018-10-27 04:21:44] 'netpositive' has started the backup! [2018-10-27 04:21:44] Marking backup as running... [2018-10-27 04:21:44] Making sure '/var/www/discourse/tmp/backups/default/2018-10-27-042144' exists... [2018-10-27 04:21:44] Making sure '/var/www/discourse/public/backups/default' exists... [2018-10-27 04:21:44] Pausing sidekiq... [2018-10-27 04:21:44] Waiting for sidekiq to finish running jobs... [2018-10-27 04:21:44] Dumping the public schema of the database... [2018-10-27 04:21:44] pg_dump: [archiver (db)] connection to database "discourse" failed: could not connect to server: Connection refused [2018-10-27 04:21:44] Is the server running on host "172.17.0.1" and accepting [2018-10-27 04:21:44] TCP/IP connections on port 5432? [2018-10-27 04:21:44] EXCEPTION: pg_dump failed [2018-10-27 04:21:44] /var/www/discourse/lib/backup_restore/backuper.rb:172:in `dump_public_schema' /var/www/discourse/lib/backup_restore/backuper.rb:37:in `run' /var/www/discourse/lib/backup_restore/backup_restore.rb:167:in `block in start!' /var/www/discourse/lib/backup_restore/backup_restore.rb:164:in `fork' /var/www/discourse/lib/backup_restore/backup_restore.rb:164:in `start!' /var/www/discourse/lib/backup_restore/backup_restore.rb:18:in `backup!' /var/www/discourse/app/controllers/admin/backups_controller.rb:32:in `create' [2018-10-27 04:21:44] Notifying 'netpositive' of the end of the backup... [2018-10-27 04:21:53] Removing old backups... The postgresql 9.6 server is running outside the container and it’s accessible on port 5433. If I manually run the pg_dump command inside the container then it works flawlessly and I haven’t detected any other DB related issue in Discourse. Any idea why does it take wrong config settings? :confused: Thanks a lot!

Oh, I totally missed the different ports in your first post. :sweat_smile: It looks like backups use the default PostgreSQL port unless specified otherwise. I think you might be able to override it by configuring DISCOURSE_DB_BACKUP_PORT: 5433 in your container config.

[image] netpositive: postgresql 9.6 server I have a feeling this may be your problem discourse/docs/INSTALL.md at main · discourse/discourse · GitHub Software Requirements Postgres 10+

No, Discourse works still fine with 9.6 and it definitely should not get a connection refused because of the Pg version.

Thanks, somehow I’ve overlooked that. Now on postgresql 10 but the problem still persists :roll_eyes:

[image] netpositive: If I manually run the pg_dump command inside the container then it works flawlessly You could try running the same command Discourse uses for dumps. You can get it by executing the following inside the container: rails c BackupRestore::Backuper.new(Discourse.system_user.id, {}).send(:pg_dump_command)

It doesn’t work because the port 5432 is used instead of 5433 :frowning: [1] pry(main)> BackupRestore::Backuper.new(Discourse.system_user.id, {}).send(:pg_dump_command) => "PGPASSWORD='redacted' pg_dump --schema=public --file='/var/www/discourse/tmp/backups/default/2018-10-29-184350/dump.sql.gz' --no-owner --no-privileges --verbose --compress=4 --host=172.17.0.1 --port=5432 --username=discourse discourse"

Thank you, that setting did the trick :relaxed:

DB backup config error

Product Support Self-hosting

netpositive October 31, 2018, 5:41am 8

Thank you, that setting did the trick

Topic		Replies	Views
Failing backups Product Support	3	610	April 3, 2018
Error when migrating to an external postgres instance Self-hosting	11	665	January 18, 2024
Pg_dump backup failure for remote pgsql - port & version differentials; what options exist? Self-hosting	6	3574	August 20, 2022
Backup has failed. pg_dump: server version: 9.6.3; pg_dump version: 9.5.10 Development	7	9214	December 2, 2017
Backup failed due to PG/SQL errors Self-hosting	9	2551	December 17, 2020

DB backup config error

Related topics