Posts made by ntnsndr

@girish Will do!

@ntnsndr Update: I just created a test campaign for 500 dummy users, and it appears to be working normally, arriving in my x+[1-500]@gmail.com inbox within about 30 minutes of initiating the campaign. In order to achieve this, I did the following in Cloudron:

• Raised Listmonk memory limit to 512
• Raised postgresql memory limit to 2 GB

And in Listmonk SMTP settings:

• Max connections to 25
• Idle timeout to 2m, Wait timeout to 5m

I am still getting the "AxiosError: Network Error" on any analytics page on Listmonk.

Note that in the past I have been able to send campaigns to large lists without changing any of these default settings, but my server has been under increased load lately.

I would appreciate any further guidance on best practices from @girish or anyone else. Thanks!

Update: two days later, this setup did NOT work for my 900 person email list. I got an even lower view count than before, with more errors.

@ntnsndr I wonder what can be done to prevent this kind of thing for future users—different Cloudron configuration settings perhaps?

Following this thread I raised the timeout period and got at least some promising results. I haven't tested at large scale. But I'm curious if others have any best practices suggestions for these settings.

Analytics are still not loading, even with memory raised on the db and the app.

@girish darn, the decline is continuing:

Is there anything else I can do? Here is the error I'm seeing in the logs, over and over:

manager.go:392: error sending message in campaign March 2023 - Sacred Stacks: subscriber f3739c80-1055-44a4-9636-476a5c3b50a8: timed out waiting for free conn in pool

@girish Okay thanks—just bumped up the memory and will try again

@girish Thanks—I'll try that. Earlier, I did actually see the view count rise from test emails I sent myself, so that seems to be working.

One other thing I'm noticing is an "AxiosError: Network Error" when I try to view the analytics on a campaign. Any idea what might be causing that?

I'm having a strange experience when using Listmonk (which in general I love!). The number of recorded views on my campaigns has been significantly declining. I can't figure out why this might be (other than lousy content?). Can anyone think of a way I could diagnose this issue?

@girish Thanks—seems to be working.

Any further thoughts on this? It isn't clear to me when the cutoff is for tokens being old. But my main Cloudron token is 4 years old. If I regenerate it in DigitalOcean, what do I do to update it on Cloudron? Just in Domains and Certs, or somewhere else?

What's also weird is that in the DokuWiki media backend, not even the default DW images show (so maybe the problem is with the whole folder?):

Never mind, the permissions are all the same. And the file appears in the Cloudron file manager. But on the wiki it appears broken.

Sorry for the delay. No, no success. I did upload them, but the logo.png image is not loading. See here: https://socialcoopwiki.medlab.host/

It also doesn't appear in the backend. Could it be that the file system permissions are not set properly?

I just saw that DigitalOcean is changing its API token management. Is this something we have to act on if we use DigitalOcean API access for Cloudron?

I've just deployed a new DokuWiki, and none of the images I have uploaded are appearing on the site. Could this be some kind of permissions issue? I'm not sure how to begin diagnosing this. But let me know if anything comes to mind!

Thanks for your help.

@nebulon Thanks for this. I think it would be great to have a "download archive" button on a per-app backup. That, or to have a kind of "cold storage" archive where the archived apps are no longer on the app menu.

@girish Ah yes—working now. Thank you!

@girish And findings? And if I need to downgrade, how do I do that?

I just saw there was an "update available," and I updated, but it still is "not responding."

I am having this issue as well. What file is the error in? I can't tell.

@jdaviescoates Me too!

@girish Just tried again and got these logs:

Jan 18 14:51:50 172.18.0.1 - - [18/Jan/2022:21:51:50 +0000] "GET /about HTTP/1.1" 301 5 "-" "Mozilla (CloudronHealth)"
Jan 18 14:51:53 172.18.0.1 - - [18/Jan/2022:21:51:53 +0000] "GET /packs/js/public-197b4c6c11335d189f0d.chunk.js.map HTTP/1.1" 200 47869 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:53 172.18.0.1 - - [18/Jan/2022:21:51:53 +0000] "GET /packs/js/locale_en-b2cbaebad015671e1ba8.chunk.js.map HTTP/1.1" 200 556 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:53 172.18.0.1 - - [18/Jan/2022:21:51:53 +0000] "GET /packs/js/common-3869dfcbde1ff39e38cf.js.map HTTP/1.1" 200 3657773 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:53 172.18.0.1 - - [18/Jan/2022:21:51:53 +0000] "GET /packs/css/default-4fb6b0ab.chunk.css.map HTTP/1.1" 200 736008 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] method=POST path=/auth/sign_in format=html controller=Auth::SessionsController action=create status=422 error='ActiveRecord::RecordInvalid: Validation failed: Username must contain only letters, numbers and underscores' duration=66.27 view=0.00 db=1.48
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80]
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] ActiveRecord::RecordInvalid (Validation failed: Username must contain only letters, numbers and underscores):
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80]
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] app/models/concerns/ldap_authenticable.rb:29:in `ldap_get_user'
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] app/models/concerns/ldap_authenticable.rb:12:in `authenticate_with_ldap'
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] app/controllers/auth/sessions_controller.rb:66:in `find_user_from_params'
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] app/controllers/auth/sessions_controller.rb:59:in `find_user'
Jan 18 14:51:55 [f0cfd5d6-782a-40ba-8fa3-9fd4e28e3e80] app/controllers/concerns/sign_in_token_authentication_concern.rb:11:in `sign_in_token_required?'
Jan 18 14:51:55 172.18.0.1 - - [18/Jan/2022:21:51:55 +0000] "POST /auth/sign_in HTTP/1.1" 422 0 "https://ism.mediastudies.social/auth/sign_in" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:56 172.18.0.1 - - [18/Jan/2022:21:51:56 +0000] "GET /sw.js HTTP/1.1" 304 0 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0"
Jan 18 14:51:58 verb Subscribe timeline:access_token:74
Jan 18 14:51:59 verb d5fcdc17-4215-46df-8a26-e9c20735496e Starting stream from timeline:107616470286852558 for 107616470286852558
Jan 18 14:51:59 verb Subscribe timeline:107616470286852558
Jan 18 14:51:59 verb d5fcdc17-4215-46df-8a26-e9c20735496e Starting stream from timeline:public:local for 107616470286852558
Jan 18 14:51:59 [39ce2b48-7679-4a4a-a29b-7a70a3b4201c] method=GET path=/api/v1/timelines/home format=html controller=Api::V1::Timelines::HomeController action=show status=200 duration=4.32 view=0.34 db=0.46

This makes it look like the error might arise from the use of a dash in the username. I guess Mastodon does not know how to handle the issue; since Cloudron handles the username creation, and doesn't catch that problem, it passes the problem to Mastodon. Perhaps Cloudron should automatically cleanse usernames for Mastodon, but that would require informing users so they follow that convention.

@girish Update: I was just able to impersonate one of the users having this problem. When I tried to log in on Mastodon, I also got a blank white screen.

In the browser (Linux Firefox), there are no errors in the Console. Here are the only log entries that came up around this time:

Jan 18 14:43:09 15:M 18 Jan 2022 21:43:09.128 * Background saving terminated with success
Jan 18 14:46:50 172.18.0.1 - - [18/Jan/2022:21:46:50 +0000] "GET /about HTTP/1.1" 301 5 "-" "Mozilla (CloudronHealth)"
Jan 18 14:47:00 172.18.0.1 - - [18/Jan/2022:21:47:00 +0000] "GET /about HTTP/1.1" 301 5 "-" "Mozilla (CloudronHealth)"
Jan 18 14:47:07 [e71d8c61-e3b5-4d23-bedb-8c847ec483b2] method=GET path=/api/v1/trends format=html controller=Api::V1::TrendsController action=index status=200 duration=3.13 view=0.38 db=0.41

View Source returns an empty document.

@girish Good questions, thanks!

• The error is strange. Some students are getting this error exclusively, on various browsers, and some aren't getting it at all. So it seems like it might be tied to login credentials somehow.
• The Event Log in Mastodon is showing students who are getting this error as logging in. So no help there. But I can send you the Mastodon logs if you like; I don't see anything there that seems helpful.

Good point about the browser console. I can try to get a student who is having this problem to help me with that.

I'm using Cloudron to manage a Mastodon instance in a course with 260 students. Most of them are logging in without issue, but some are reporting that they are getting a blank white screen when they attempt to log in using their Cloudron credentials.

• I saw a student do this on Safari, as well as on a mobile iPhone app.
• I saw a screenshot of a student who got a 422 error from logging in on Chrome.

Does anyone have any experience with this? Or any guesses about why it might be occurring for some users and not others?

Thanks very much for considering.

I am using Cloudron to manage a university class with 250+ students. I wonder if there is a way to easily add all the students to a specific group, without having to add them manually one by one. For instance, can I:

• Send them a magic link that they can use to create a new account in a given group?
• Import them via a csv file or the like?

Thanks for your help.

@brutalbirdie Mostly... I am in contact with them, and they said that they unblocked it from their end. But perhaps they did not do so correctly.

I'm using Cloudron on Linode, too, and I can't get outbound email working—despite getting some assistance from the Linode people—though now they're saying I need to handle the configuration issues on my end. Here's what I'm still seeing:

I tried using ufw here to open up 25/tcp:
https://www.linode.com/community/questions/19219/how-do-i-open-a-port-in-my-linodes-firewall

Can you think of anything else I can be doing?

@murgero Thanks—yes, it does support Webhooks, as well as Git Hooks. But any suggestions how to handle the authentication process that GitHub Pages requires in a script?

Hi! I'm trying to replicate a GitHub/Lab Pages setup using the combo of Gogs and GitHub Pages on Cloudron. Basically the goal is that anytime a commit occurs on Gogs or local changes are pushed to it, I want it to deploy to the GitHub Pages site. I also want to do this in a way that doesn't expose the credentials to GitHub Pages in a git config data file.

My assumption is that, if this is possible, it would be through Git Hooks on Gogs. But I don't see how best to include login credentials there. I'm not the best at shell scripts. Can anyone help me with this?

Thank you so much.

Agreed! Any progress? It seems like it could be better than Commento.

Has anyone gotten a decent editor for DW working on Cloudron? I really can't stand DW syntax, and the major markdown plugin doesn't work for the current version of DW. I've tried installing CKGEdit Plugin but on Cloudron it doesn't work; the editor never appears, and when I try to find it inserts a weird "false" into the page I was trying to edit.

Any advice?

@girish Thanks so much for this! Just what I needed. I'm working on creating an example repo that I can use for future. I think I fully de-cloudron-ed it... Am I right to leave in .gitlab-ci.yml this part?

default:
image: cloudron/docs-ci@sha256:4a63ad48a6fbbd168828769b0d796b511d09d173ab5dd64368338997164aa4ab

Here's the repo: https://gitlab.com/medlabboulder/docs

Let me know if there's anything else I should do. But for now it deploys!

@3wordchant This is really exciting. Thank you for sharing this project!

Has anyone tried to push an MKDocs site from GitLab to a Cloudron app, either Surfer or the GitHub Pages app? I would love guidance on how to do that. It'd be great to manage the site on GitLab but have it push to Cloudron.

Thanks!

I would love to see this as well. Am experimenting with Minetest at my lab. Is there any chance we could try this?

Thanks for raising this question, @jdaviescoates.

I am personally not sure open-sourcing is critical here, as I think the first beneficiaries would be big cloud platforms (AWS, etc) that would then be able to host it and cut out any income for the developers. The current source available arrangement strikes me as probably necessary and appropriate.

I'm increasingly convinced that OSS as such is broken as a means of protecting against corporate exploitation, and it should not be celebrated as an end in itself. Based on my conversations with @girish, I think the single most important fact about Cloudron is that the company is bootstrapping (based on our subscription fees) and not seeking an exit. As long as that's the case, I think the community should support their self-defense through licensing.

Rather than fixating on licensing, it might be more relevant to all of us to discuss the possibility of an "exit to community" for Cloudron, in which ultimately the company we pay into becomes owned by—and accountable to—the people who rely on it. This could help ensure that the company we're paying into, and that is stewarding the code we depend on, doesn't get captured by forces beyond our control.