Is latest-release+105 and v2.4.0beta4+105 the same?

I have a discourse forum running.

I have created a tested machine and installed discourse and upgraded.

I have upgraded the forum machine too.

They should be running the same discourse version.

But when I look at versions in the admin/upgrade page, the forum machine says it is running latest-release+105 and the test machine v2.40beta4+105.
Both of them say that they are up to date.

It seems they are using different git repos, but I don’t remember to have changed anything in the forum machine.

What is going on?

I am seeking for differences in versions because then I try to retore a backup from the forum machine to the test machine I get an error.

Both machines have installed the same plugins and I have upgraded all to the last version.

The error I get when trying the restore is this:

[2019-09-24 20:11:58] [STARTED]
[2019-09-24 20:11:58] 'admin' has started the restore!
[2019-09-24 20:11:58] Marking restore as running...
[2019-09-24 20:11:58] Making sure /var/www/discourse/tmp/restores/default/2019-09-24-201158 exists...
[2019-09-24 20:11:58] Downloading archive to tmp directory…
[2019-09-24 20:12:29] ALTER TABLE
[2019-09-24 20:12:29] ALTER TABLE
[2019-09-24 20:12:29] ALTER TABLE
[2019-09-24 20:12:29] Migrating the database...
[2019-09-24 20:12:32] EXCEPTION: PG::UndefinedColumn: ERROR:  column user_profiles.card_image_badge_id does not exist
LINE 1: ..._version", "user_profiles"."badge_granted_title", "user_prof...

[2019-09-24 20:12:32] /var/www/discourse/vendor/bundle/ruby/2.6.0/gems/rack-mini-profiler-1.0.2/lib/patches/db/pg.rb:72:in `async_exec_params'
/var/www/discourse/vendor/bundle/ruby/2.6.0/bin/unicorn:23:in `load'
/var/www/discourse/vendor/bundle/ruby/2.6.0/bin/unicorn:23:in `<main>'
[2019-09-24 20:12:32] Trying to rollback...
[2019-09-24 20:12:32] Rolling back...
[2019-09-24 20:12:33] Cleaning stuff up...
[2019-09-24 20:12:33] Dropping function from the discourse_functions schema
[2019-09-24 20:12:33] Removing tmp '/var/www/discourse/tmp/restores/default/2019-09-24-201158' directory...
[2019-09-24 20:12:33] Unpausing sidekiq...
[2019-09-24 20:12:33] Disabling readonly mode...
[2019-09-24 20:12:33] Marking restore as finished...
[2019-09-24 20:12:33] Notifying 'admin' of the end of the restore...

How did you installed the test machine?

By default the Discourse instances run on test-passed branch, latest-release might be the one used for development as you need to test a lot of things before those changes go to production.

Thank you for your help.
As long as I know, I have not changed the default configuration. I don’t even know how tl change the repository of installed release.

Till know, when I looked for installed version it reported v2.3.0beta2 v2.4.0beta2 and so on.

But when I have upgraded last time, it now reports latest-release in the production machine.

May it be some kind of bug in the upgraded script that changed the repo?

Strangenly enough, the test machine reports v2.4.0beta4 after upgrading.

It worries me.

We should not be running untested versions in production machine.

How may I securely return to production repos?

discourse-doctor reports that version v2.4.0beta4+105 is installed in both machines

Latest is the correct setting, all our hosting is on latest. You are crippling yourself if you run anything else, if you want to be super super conservative you can run stable but you will be months old.

Thank you, Jeff.

If latest is the usual setting for installing discourse, for me is OK.

I was surprised that having updated both computers, they show different versions (one latest-release and the other v24.0).

By the way, I don’t know where the repository is changed from which to install discourse, where can I do it?

1 Like

I think you’re both out of date? It should look like this:

This is what I get in the /admin/upgrade page of my production site:

And this is the same page in the test machine, installed following the same steps I dis in the production machine, and upgraded a the same time (none is up to date as there has been some commits since I upgraded both).

They don’t report the same (latest-release against v2.4.0beta4).

When I restore a backup from production into test machine, I get the error I mentioned previously.