Posts made by robi

@shan Why not keep using the method you have with larger number?

You may also wish to set up IP aliases on the Cloudron IP interface ex: eth0:[0-3], so all IPs go there.

@girish Looks like you need some dark mode there to help you with the strain

@ruihildt Sure, you can run it without using it yourself and support the network.

And yes, we can start with a simple page that simply says to manage this app, log in to the dashboard. Maybe add an 'app is running' JS check.

Sounds like this might install in the LAMP App.

Anyone try it?

@arshsahzad probably a typo in the .conf file.

If you can reproduce it, post the contents.

@nebulon it's headless like tor.
Either on or off and you point your browser at it.

You can make it more like the browser plugin works, which reports some basic stats as a simple web interface.

See https://community.torproject.org/relay/setup/snowflake/browser/

Take a look at the plugin and how it retrieves status data from the proxy and then generate/copy the same html page.

Maybe try creating a filter in Snappymail and see how it is different in Thunderbird.

do we need a custom role?

Or a way to decouple email from cloudron?

@JLX89 there were plans for providing per App VPN via Wireguard.

Uses node 16+ and a few config files

Here's the Dockerfile

See if it builds..

@marcusquinn said in OpenLiteSpeed Wordpress:

@robi ElasticSearch, yes: https://git.cloudron.io/vladimir.d/elasticsearch

So you point the ElasticPress plugin directly at ElasticSearch?

Or is ElasticPress a separate packaged install?

@marcusquinn said in OpenLiteSpeed Wordpress:

@robi It's based on this plugin: https://wordpress.org/plugins/elasticpress/. Plus some custom add-on plugins we wrote for our custom theme.

Ah okay, that makes it a bit simpler. Does it work across subdomains?

Do you have https://github.com/10up/ElasticPress packaged for Cloudron?

Hmm, I didn't find an app request for it..

@marcusquinn said in OpenLiteSpeed Wordpress:

(Since that was written, we reproduced what Algolia does with an in-house developed ElasticSearch alternative, also hosted on Cloudron.)

Is there a how-to on that?

@JOduMonT Seems like linking that over to /app/data should do the trick.

See if you can do that in recovery mode.

@parhelium You can:

• package it yourself
• sponsor/pay a dev to package it
• ask the original dev to package it
• create a packaging event for learning and group packaging
• get more upvotes for this post
• wait

Updated first post with more data.

@fbartels

Yeah, common problem for minimal app installs.

Container images are a good source of precompiled binaries and libs, so if you know how to pick those out of the images, that can be a light recipe.

Statically compiled binaries are a bit much, however separating the /bin/ dirs into persistent layers would also work.

@fbartels said in Webtop - Dockerised Linux Desktop in a Browser:

its biggest drawback is that one cannot easily install new software

The video clearly shows that is not the case, as you can apt get anything you like.

Tim does mention that this is more like a container you don't update via a new container image, but only via updates from within, just like the usual OS and App updates.

This of course just grows the container continuously over time.

We'd need something like what Cloudron does for our apps, separating the app from the data and being able to restore it after an image update.

Or by mounting dockerized app images as layers instead of installing apps, preserving data elsewhere.

@fbartels said in Add a Note about 'Packet Time' to the 'Backups' section in Docs:

@michaelpope i think the better word here would be "latency".

Could do both.

When is the next event?

@nebulon While the UI issue is obvious, there are missing hours in the list and missing days on disk.

A client recently changed from local to CIFS based backups to a local NAS.

Since the changes of the backup from local to local (both in a VPS and from/to NVMe targets) to an external NAS Windows share on it (several weeks ago), the backup retention policy isn't followed anymore.

After this was detected, the retention policy was manually changed/reset, hoping that this would fix the problem - but it didn't.

The result is that no backups older than few days (see screenshot below) are kept, and overall they are way too few compared to the retention policy settings:

In the above screenshot, after the "just now" backup was completed, the 4th day backup disappeared from the list. (UI issue?)

The backup policy is set to hourly, minus the 4am maintenance time window for possible app updates, and these backups are successfully done each hour as intended - though old ones get constantly removed from the list (and disk) despite the above retention settings. This has not been an issue before when the backup was done without compression to a local disk.

Space on the target NAS is not an issue, as there are 4TB free besides the existing backup of about 860GB, (about 5mins backup time).

Update 1:
This is the NAS backup directory with its content and directory sizes, and it seems to neither match the intended retention settings nor the listed backups:

Update 2:
8h later the backup list looks like this:

@jdaviescoates I read it as a contrast to WHM, which is an app, not as a CL App.

@nebulon @jdaviescoates

https://www.qwant.com/?q=WHM

@Refugee_Ranger While the default is to use Cloudflare for DNS only, as it's required to install and set it up, you can manually switch any Apps you like to proxied as long as they're on port 80/443.

Take a look at https://forum.cloudron.io/post/43487

@dylightful I hear you.. it has not been made clear yet.

I just managed to deploy wg-easy in fly.io and it's simple UI is great, doesn't need a username, and similar to our OpenVPN app, easily generates .conf files for download for the clients.

For some of the things we wanted to do with VPNs for Apps which were a lot more complex, a lot more integrations were needed, and the people who started doing those didn't manage to complete them and the chain of events stopped progress.

What we perceive being reality, this can affect much simpler things from being re-prioritized; and of course life happens.

Un/fortunately those are not blockers for Cloudron having a fast personal VPN experience via Wireguard.

As I have a bit more time this month, I may start packaging wg-easy, and if someone else is interested in lending a helping hand, many hands make short work. (Send a PM to collaborate)

@vladimir-d nice work! Thank you.

If you can, join us for the upcoming packaging events where your knowledge can be shared. (Or host your own!)

There one can learn how to overcome the current issues and more.

The next steps of polishing a new app ready for testing and packaging on Cloudron will be a good learning experience too, for folks who have found their way this far.

You're invited!

Hosting an own instance should be possible now and it's under active development.

https://github.com/internxt/drive-server

Perhaps making a good backup target too?
Or self-hosted Dropbox like own files..

@fbartels Thanks for the reminder!

@fbartels How about one or more of the Wireguard Apps?

That's another good topic, show and tell on how you do 3rd party monitoring with flies, haha. bzzzz

@roofboard said in Security Recommendations...:

So @robi I think I just got picked up by a crawler. I caught the hack in a matter of hours,

Yep, can't be lazy, as botnets are scanning the entire IP space for targets constantly.

And yes, many revisit previous active targets for more interesting exploits for a short time before moving on.

So order of operations and not skipping crucial initial steps is important

It happens. Lesson learned.

Could have been worse if you were not on Cloudron.

Managing backups like we mange mailboxes does sound like a good idea, and can have checkboxes for daily, full, etc. per app, and more.

@infogulch that's a bit odd IMO to have to find them, as they should be explicitly defined.

Perhaps make a post in App Dev section to get assistance with the solution you're working on, as a step in the right direction.

@BrutalBirdie said in Regular App Packaging Event:

So everyone could monitor their own Cloudron . . . until it goes offline and then takes Zabbix with it

Nice, we can do better, monitoring ours and each others for a few key services to avoid being notificationless

@scooke However you like!

If it's an in-person meetup or online, there is a host who is organizing and setting up the event, maybe recording it, etc.

They can choose an app they want for the event or another person like an app dev can choose one.

Or you make a deal with both to separate duties, while you try the packaging yourself with guidance from others. Fun!

The idea is to choose one we haven't done before, perhaps in a different language than before, so we cover more of the spectrum.

Just choose one, any one in the App Requests will do and move things forward.

Latest stable version 1.12.1 is out on their website.

Feels like it's time for more of these events!

Who's in?

Anyone having more events like this?

It seems @jimcavoli already got this working, see here:
https://forum.cloudron.io/topic/3678/build-deploy-to-cloudron-from-github-actions

@infogulch

One option would be to include all the Metadata as a layer for the image, making it self contained and having its own layer ID.

That way we're tagging what is needed.

JFrog did this in a similar way, see https://jfrog.com/knowledge-base/what-is-the-docker-image-layer-marker/

Much simpler than tagging files across multiple image layers, etc.

@roofboard said in Security Recommendations...:

a wordpress site got hacked...

A few questions:

Who did you piss off?

Which WP Cloudron App were you using?

Were updates enabled? For the plugins too?

What was changed during the hack?

Why is it a headache? (Other than it happening)

Have backups pre-hack to restore? Easy-peasy?

As @murgero said, Cloudron was designed to mitigate these types of things in many ways, and getting back online is much easier thanks to that.

Would not having Portainer as an App allow us to manage containers and allow for trying any new unpackaged apps straight from dockerhub or other repos?

This works well having tried the demo in Runtipi.

This would solve the single DB App needs, custom LAMP setups and all the cool new stuff we want to try.

Another option would be to have an empty Appless App where one can manually install any other app, managing things oneself.

A nifty option is using Docker-in-Docker via Nestybox where in one container you can run further docker images nested and more isolated with better security.

All these need a way to configure the ports, via App settings or during install, so it's accessible via the system Nginx reverse proxy.

One app solving many requests, yes?

An example config of WG-Easy deployment from:
https://github.com/WeeJeWel/wg-easy/wiki/Using-WireGuard-Easy-with-nginx-SSL

docker-compose.yml:

version: "3.8"

services:
  wg-easy:
    environment:
      # ⚠️ Change the server's hostname (clients will connect to):
      - WG_HOST=wg-easy.myhomelab.com

      # ⚠️ Change the Web UI Password:
      - PASSWORD=foobar123
    image: weejewel/wg-easy
    container_name: wg-easy
    hostname: wg-easy
    volumes:
      - ~/.wg-easy:/etc/wireguard
    ports:
      - "51820:51820/udp"
    restart: unless-stopped
    cap_add:
      - NET_ADMIN
      - SYS_MODULE
    sysctls:
      - net.ipv4.ip_forward=1
      - net.ipv4.conf.all.src_valid_mark=1

  nginx:
    image: weejewel/nginx-with-certbot
    container_name: nginx
    hostname: nginx
    ports:
      - "80:80/tcp"
      - "443:443/tcp"
    volumes:
      - ~/.nginx/servers/:/etc/nginx/servers/
      - ./.nginx/letsencrypt/:/etc/letsencrypt/

~/.nginx/servers/wg-easy.conf:

server {
    server_name `⚠️wg-easy.myhomelab.com`;

    location / {
        proxy_pass http://wg-easy:51821/;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "Upgrade";
        proxy_set_header Host $host;
    }
}

@braginini Would you be open to helping package this for Cloudron?

@staff does the WP dev base image contain PHP 8.1? If not, can you update it?
That would simplify the issue here.

@kris that's odd about FTP, as all my WP installs never ask for such, and just install plugins with a click.

Perhaps your permissions are wrong so it prompts for alternate ways?

You can always install things manually through the Cloudron built in File Manager or the Terminal. It's in the App config section of the Cloudron UI (gear icon).

@nebulon said in Simple Database Only Apps (Mysql, Postgress, Redis, MongoDB...):

Maybe we can start collecting the various databases as normal app request in the forum to see which ones to target first?

There's only a handful, just do them all, no?
Best for Cloudron and the community.

I think we need the Wireguard VPN apps first, so we can securely connect to the the private DBs.

Btw, you can already use the DBs that other apps use, such as LAMP.

Once the App is running you can find the connection info in the output App terminal env command.

Nice find.

So this would encompass the features of change detection?

@scooke yes, natural massage

Where was it?

Have you searched the forum?

I found this.

@allanbowe that would be the package version which @staff would have to modify.

There is also the Nginx reverse proxy for the app in the host OS.

perhaps this error response is from nginx?

@ruihildt said in Snoop Forms — fully FOSS code/nocode form builder:

Seems trivial to package. But it's still not stable.

Perfect for the unstable section of the appstore.

@marcusquinn Perhaps their dockerfiles help with packaging here or at least easier adoption from pure code sources.

@potemkin_ai I hear you, and the frustration behind this, however you did manage to get sound working after fixing the port issue.

Hence the topic is marked solved.

The other issues you are having are not related to Cloudron or it's supported use. That part is up to you, right?

ISPs are highly complex stacks with many moving parts.

From the networking equipment, firewall rules and other protection and network address translation technologies, each is different.

Then comes the virtualization stack and configuration, and another layer of networking specific to that inner stack.

Then you get into the OS install and Cloudron managing itself.

I hope you can appreciate that this forum helped you solve a self caused user error.

I don't agree with your conclusions, hence I am replying here to help you see where to look, however my time is limited and free consulting has its limits.

Inconvenience or not, your unmentioned choice of ISP and many nested stack isn't working for you as you expect. Hence the suggestion to try another ISP, many of which offer free trials even.

@potemkin_ai said in No sound in Jitsi...:

apologies, not sure when I blamed anyone

It's not about blaming people or that you did so, it's about finding fault with systems that work for nearly everyone else.

You're posting here, in a Cloudron forum, where Jitsi, docker and the entire stack just works.

You had a "reading" / port issue, so this topic is solved. Right?

Perhaps it's time you look at a different VPS provider, or ask in another topic which ones have no issues with having Jitsi installed.