Cloudron Forum

I needed to comment out a number of lines in the site_settings.yml file before Discourse would finish starting up.

Specifically, I commented out the following:

• users.min_username_length: min line and max line
• users.max_username_length: min and max
• basic.categories_topics: min and max

After doing this, everything appears to be running smoothly!

It's wonderful to see the Cloudron packaging updated so quickly after the Discourse release! I wanted to share a few notes on my upgrade to v3.1, in case it helps anyone else.

First, I had to remove login.invite_only.validator (which was previously set to "EnableInviteOnlyValidator") from site_settings.yml in order to get the server to start after updating.

Second, one of the featured additions in the 3.1 announcement was the modular Discourse AI plugin. I was interested to test it out, but was unsuccessful in trying it. After adding the plugin and restarting the server, I got caught on this:

PG::UndefinedFile: ERROR: could not open extension control file "/usr/share/postgresql/14/extension/vector.control": No such file or directory

Looks like there is a Postgresql extension that needs to be installed/configured to use this plugin.

@girish Two ideas based on this experience for small additions to https://docs.cloudron.io/apps/discourse/

1. Would be great to mention Safe Mode as a useful debugging tool on this page.
2. Would be helpful to add a section on how to uninstall plugins. When I searched online, everything pointed to removing an entry from app.yml, which doesn't exist in the Cloudron setup. I'd imagine it's a modified version of the current section on installing plugins: removing a directory and perhaps re-running plugin:install_all_gems but that's just a guess.

I originally posted this issue upstream at https://community.n8n.io/t/all-workflows-deactivated/49352/1, but given that at least one other person who reported the same issue was also running on Cloudron, I think it’s worth flagging here.

After updating to n8n 1.47.1, I started seeing this error:

WorkflowActivationError:

There was a problem activating the workflow: “crypto is not defined”

Looking through recent issues, this one caught my eye given Cloudron login: https://github.com/linkwarden/linkwarden/issues/1422

@girish It did resolve the issue! Currently running n8n@1.48.3 with no problems. Thank you for the upstream coordination!

For me: Cloudron login, one admin user, public disabled. API calls failed after update, which tracks with the database being reset.

Same. Thank you again to the Cloudron team for the excellent automated backups implementation. One click to restore the snapshot taken immediately before the bad update, and everything lost is back.

@girish The site broke a few weeks ago, so the delay was just on my end in looking into it. After hand-updating a long list of deprecated settings, the app starts again. For reference, here is the full list of changes that I needed to make in site_settings.yml:

use_admin_ip_whitelist => use_admin_ip_allowlist
email_domains_blacklist => blocked_email_domains
email_domains_whitelist => allowed_email_domains
unicode_username_character_whitelist => allowed_unicode_username_characters
user_website_domains_whitelist => allowed_user_website_domains
whitelisted_link_domains => allowed_link_domains
embed_whitelist_selector => allowed_embed_selectors
auto_generated_whitelist => auto_generated_allowlist
attachment_content_type_blacklist => blocked_attachment_content_types
SiteSetting.enable_sso => SiteSetting.enable_discourse_connect
attachment_filename_blacklist => blocked_attachment_filenames
blacklist_ip_blocks => blocked_ip_blocks
whitelist_internal_hosts => allowed_internal_hosts
whitelisted_crawler_user_agents => allowed_crawler_user_agents
blacklisted_crawler_user_agents => blocked_crawler_user_agents
onebox_domains_blacklist => blocked_onebox_domains
inline_onebox_domains_whitelist => allowed_inline_onebox_domains
white_listed_spam_host_domains => allowed_spam_host_domains
embed_blacklist_selector => blocked_embed_selectors
embed_classname_whitelist => allowed_embed_classnames
---
SiteSetting.sso_overrides_email => SiteSetting.auth_overrides_email 
SiteSetting.sso_overrides_avatar => SiteSetting.discourse_connect_overrides_avatar
SiteSetting.enable_category_group_review => SiteSetting.enable_category_group_moderation
SiteSetting.newuser_max_images => SiteSetting.newuser_max_embedded_media

After the last update, I'm no longer able to view my forum. Restarting app didn't help.

When I load the app in browser, I see a loading indicator bouncing indefinitely. The browser console suggests that there is an issue with Ember (v5.5.0):

I took a look in my Cloudron app logs, and don't see any notable errors. Saw a few iterations of these two messages while on the loading screen, which I suspect is normal/expected, but wanted to mention just in case.

Feb 07 08:45:00 Redirected to <...>/login
Feb 07 08:45:00 Filter chain halted as :redirect_to_login_if_required rendered or redirected

(This Discourse is configured to put all functionality behind a login.)

I just discovered Safe Mode, and that helped me narrow the source of the issue to either the theme and theme components that I'm using. Next step is tracking that down, but this helped considerably.

Wanted to share in case others run into the same issue! Visit <discourse_app>/safe-mode to toggle off three different potential sources of issues.