• knfrmity
    link
    fedilink
    arrow-up
    9
    ·
    1 year ago

    Looks like a cool app but it’s super unpractical and hacky to self host and you can’t even use a bunch of features when it’s deployed that way.

    • krash@lemmy.ml
      link
      fedilink
      arrow-up
      1
      ·
      1 year ago

      Can you elaborate on your experiences so far? What’s required in order to selfhost it, and what features will be missing…?

      • knfrmity
        link
        fedilink
        arrow-up
        4
        ·
        1 year ago

        From their blog post (linked to by the docs page) about self-hosting:

        The following Omnivore features will not be included in this minimal Omnivore setup:
        - The web app (we will use the iOS app from the AppStore as our client)
        - Search of PDFs
        - Saving URLs instead of pages (more on this below)
        -Receiving newsletters via email
        - Text to speech
        

        Not only that, they use a non-self hosted elasticsearch provider.

        Their example docker-compose file in the repo has no less than four containers defined, not including the database server, and you have to build them all yourself, so it’s more of a local dev environment type deployment rather than production.

        Here’s their “make self-hosting more practical” Github issue, coming on two years old with no progress: https://github.com/omnivore-app/omnivore/issues/25

        All of that was more than enough for me to not even bother to try to deploy my own instance. I manage with Wallabag for now, it’s not the greatest implementation either but at least it can be self-hosted. Omnivore looks slick but the backend just doesn’t keep up.

        • krash@lemmy.ml
          link
          fedilink
          arrow-up
          2
          ·
          1 year ago

          Thanks for this. I think this is also an example of a opensource software that is selfhostable, but is intended for a different audience. I think Zammad, Monicahq etc fall under this category. I suppose one would need a solution with an entirely different architecture that’s aimed for selfhosters, rather than hope that omnivore becomes easier to selfhost.

          • knfrmity
            link
            fedilink
            arrow-up
            1
            ·
            1 year ago

            As far as I can tell this just fixes some relatively minor issues the contributor was experiencing deploying the hacky self-host stack on his Kubernetes home lab or private server. It doesn’t bring anything new to the table, and I’m not sure that a full on Kubernetes or similar distributed swarm deployment is really what the average self-hoster needs or even wants.

            It could just be that Omnivore tries to do too much for it to be feasible to self host. It was also conceptualized with certain third-party backend services in mind which makes it tricky to adapt.

            Maybe I’m asking for too much, but I would be looking for a two service stack, one for the app and one database service. The current and forseeable future state of Omnivore is four backend services excluding the database, and like I already pointed out you’re not even getting the full feature set.