/e/ Cloud - selfhosting

Hello !
je viens de proposer à la team de /e/ de packager leur solution de cloud auto-hebergé pour Yunohost.
Leur solution est vraiment chouette et ça vaudrait le coup, à mon sens (à commencer pour mon propre usage perso, et j’imagine celles et ceux dans ma situation), de voir les efforts de /e/ et de Yunohost se rejoindre sur cette solution.
Mon issue gitlab est ici : https://gitlab.e.foundation/e/infra/ecloud-selfhosting/issues/76

On va voir ce que ça donne :slight_smile:
Mais si certain-e-s d’entre vous, plus opérationnel-le-s que moi sur le packaging d’app sont partant-e-s pour y contribuer, ce serait chouette !

merci !


Hello !
I just posted an issue on /e/ gitlab in order to make their self-hosted cloud solution a packaged app for Yunohost.
I think it would be great (at least for people like me) to see /e/ and Yunohost in a common effort to bring this cloud solution as a Yunohost app.
link to my issue on gitlab is here : https://gitlab.e.foundation/e/infra/ecloud-selfhosting/issues/76

Let’s see what’s coming from this. :slight_smile:
But basically, if any of you, more able than me to package an app for Yunohost, is available to help them get this done properly, it would be great!

thanks!

4 Likes

Hmyea that’s something which was already suggested in https://forum.yunohost.org/t/yunohost-et-e/8787and we did not really continue the discussion because I guess we were too busy with many things …

I still don’t know much about /e/ … my understanding is that it’s a smartphone OS that is able to connect to a cloud of services managed by the /e/ people but there’s a strong wish that you should be able to self host all these services ? And that’s apparently what https://gitlab.e.foundation/e/infra/ecloud-selfhosting helps doing ?

With that said it’s really not clear to me if this is really something that is packageable to YunoHost : it sounds way too “meta” ? I mean this in the sense that it sounds like the repository is not about setting one specific web app, but more about setting up a whole bunch of services including a mail stack for example.

So I really doubt that you can have this as a Yunohost app, because it looks very much like it is itself an ecosystem, and you can’t have two very different ecosystem on the same machine … (even though the software are probably very similar, still, they are probably configured differently)

What I think would make more sense is to make Yunohost compatible with /e/, or viceversa make /e/ compatible with a Yunohost server … I have no idea what it implies though. But that would make perfect sense as we’re regularly thinking about how it’s annoying to install and configure every smartphone apps by hand, and how instead an integration between personal servers with smartphones would really contribute in adding meaningfulness and smoothness to the whole degooglization process to the general public !

In fact, I’m really afraid that this ecloud-selfhosting repository is just reinventing the same wheel as YunoHost, FreedomBox, Sandstorm, Cloudron and other friendly self-hosting distributions … But I don’t know if the /e/ folks realize how time consuming it’s going to be on the long-term to manage the various pieces of config, maintaining/upgrading them and the base system (Debian in our case), app packaging (and upgrades!), DNS configuration and port forwarding tutorials, the mail stack nightmare, security considerations, diagnosis, and making sure that all this is as easy as it can be for the end user (and certainly YunoHost itself is still only easy for tech-savvy folks…).

I don’t know how much thinking went into this self-hosted /e/ cloud or what its true goal is, but imho it’s either going to remain something that only only a few nerds-with-a-lot-of-free-time-and-energy will be able to setup and effectively maintain in the long term, - or - if they truly want to build a user-friendly self-hosting solution (and my understanding is that it’s not the priority of /e/ but I might be mistaking or it could evolve) they are going to have to invest several people full-time on it.

Instead, it would be really nice to see them choose an existing self-hosting project and trying to create compatibility with it… Of course here we’re on YunoHost forum, but any solution they choose among YunoHost, Freedombox or Sandstorm or whatever would make more sense in the long term than reinventing this wheel themselves… Though I do understand that in the short-run, writing your very own stack means having quick control over all the tiny details and therefore allow to quickly make progress.

Anyways … With that said, personally I’m already way too busy with everything going on in YunoHost itself and other related projects. But there’s really something meaningful to work on here if anybody is interested … Like “filling the gap between /e/ and YunoHost” … installing a Yunohost, installing a smartphone with /e/, and trying to see if and how you can connect pieces of one to the other, or what it would take to make it work. I guess if we had a list specific details to implement in YunoHost to make it compatible with /e/ we could work on it as long as it’s compatible with the philosophy of the project :stuck_out_tongue_winking_eye:.

Sorry for the wall of text :man_shrugging:

2 Likes

That’s hell of an answer :ok_hand:t2::rofl:

Thank you for all the precise points you made.
And now I understand way more how those two projects can cooperate at some point, or at least, be helpful for each others.

I definitely would like to administrate my self-hosted /e/ cloud services the way I do with my yunohost.

And yes, I believe in some kind of all-inclusive service to be more efficient that separated apps when it comes to ungoogle the way non-geeks use their smartphones, email services, clouds and everything.

1 Like

Yes, in fact it can’t be stressed enough but I believe one thing we (the FOSS community) are failing is realizing how much of a killer-feature it is the way Google is able to deliver a set of unified integrated services … I don’t know exactly how that works precisely, but I’m guessing that for example you can easily attach a file from Drive to a mail in GMail ? And open said file in Google Docs ? Also their services have a unified esthetic.

That’s also why Framasoft’s Degooglize campaign was succesful, they did not just installed a bunch of services and wrapped them with some clumsy tape and call it a day : they worked hard on the feel and esthetic and integration on their platform and set of services so that everything looks simple and consistent. Imho the CHATONS / LibreHoster project need to acknowledge and work on this to be succesful especially now that Framasoft’s plan is to close services.

Unfortunately in YunoHost we “only” have the SSO aspect - which is already quite nice - but apart from this the application themselves all have different looks and I’m guessing you can’t attach a file from Nextcloud to an email in Rainloop …

Nextcloud is much better at this since they have a unified / integrated environment of apps which allow this (at least in theory? never tried all this). But the major drawbac is that they have to throw away existing FOSS software and reinvent the wheel in their own framework (e.g. a task manager, a mail client, a social app, …). Sandstorm tried something different with their ‘grain’ system but that was not really straightforward either.

/e/ is also apparently putting quite a lot of effort in creating a consistent and well-integrated ecosystem on smartphone, so that’s very nice to see :+1:

The dream would be to see all those major FOSS software (Nextcloud, Rainloop, Collabora, Matrix, Wekan …) which cover the most fundamental internet services to work much closely together on being able to interface and integrate with each other, both in terms of features and esthetic … :confused:

That’s kinda related to the fact that we (the FOSS community) have a hard time attracing designers/graphisms and having development process where they can easily contribute on the “not-code” stuff

Anyway I’m wandering a bit far away from the original topic :stuck_out_tongue:

1 Like

But this wandering could be helpful to both communities imo.

I guess /e/ is trying its best to get the best out of those open-source projects to build an entire unified ecosystem, from laptops OS to cloud services through, of course, smartphones ROM.

It is to me, so far, the initiative that understood the most the needs of people like me (no that geek, but not that noob) and our parents (non-geeks at all).

That’s why I’m trying my best to connect both communities and knowledge of Yuno and /e/.

1 Like

Hello guys, very very interesting topic here. Thanks @Aleks for your long messages. I agree with what is said here.
I can tell you a few things, as I have yunohost server, and an /e/ phone :wink: and moreover I sometimes contribute in translating /e/ contents to french (website, newsletter, and one day I hope apps).

So actually I do connect everything manually :

  • Nextcloud
  • Nectcloud notes
  • Nextcloud Caldav and cardav
  • Nextcloud PhoneTrack/Maps
  • Nextcloud SMS
  • Emails in the K9-fork app
  • Password manager (bitwarden actually, previously keepass+keeweb)
  • Tiny tiny RSS

It is manual but it works.
I think one first step would be to define a kind of “auto discover feature”, like the .well-knows addresses for CalDav and CarDav or like Microsoft uses for exchange in cloud servers, so that an /e/ phone could find with just on URL (and obviously username and password) where are mails servers, nextcloud stuff, contacts and calendars, etc.

I have no time (and probably not enough skills) actually to do that, but I would be happy to see this being done. :stuck_out_tongue:

Exactly the same here.