Urgent: site down after rebuild

Support
1

Hi,

I urgently need some help - I did an upgrade from /admin/upgrade today, but the process never finished. I then restarted Discourse and tried a ./launcher rebuild app from the command line, but I keep getting the following error:

	SET first_unread_at = u.created_at
	FROM users u
	WHERE u.id = us.user_id
	
rake aborted!
StandardError: An error has occurred, all later migrations canceled:

Index name 'index_topics_on_updated_at_public' on table 'topics' already exists

The problem remains after disabling all plugins. Our site is down right now, I’d appreciate some help to diagnose and fix this. Thanks!

Update: I found that I could restart Discourse with ./launcher start app, but I’m not entirely sure of the current state of my instance.

Update: After starting the app, I went into psql and deleted the index that was mentioned in the error. After a rebuild, the exact same error came up again.

2

This should fix it, but you will need to wait till tests passed

https://github.com/discourse/discourse/commit/8f4527c1f1861e5038bf6104fc86e6c8308c03c4

if you nuked the index though it should surely pass, but maybe you did not nuke it properly… I don’t know.

5 Likes
3

:+1:thanks Sam!

maybe you did not nuke it properly

That’s very possible, Postgres is still new to me :slight_smile:

wait till tests passed

How will I know when that is?

1 Like
4

When the commit appears in this list:

https://github.com/discourse/discourse/commits/tests-passed

3 Likes
5

It’s in tests-passed

6 Likes
6

Thanks to @pmusaraj our faster build times are paying off :smiling_face_with_three_hearts:

5 Likes
7

My rebuild just passed the part where it creates the index. Thanks, gentlemen!

5 Likes