So I set up a self-hosted lemmy instance to play around with but I’m a bit confused by how the federation system is supposed to work.

I followed these instructions to get setup:

https://join-lemmy.org/docs/administration/install_docker.html

I then found that I was having issues deploying the lemmy proxy as it could not find the nginx config. I was able to resolve that issue by adding the nginx internal config from the git to my instance like so:

wget https://raw.githubusercontent.com/LemmyNet/lemmy-ansible/main/templates/nginx_internal.conf

I noticed that I couldn’t search for any instances from my personal instance, so in the “Allowed Instances” in the server settings on the web UI I added: lemmy.ml lemmy.world programming.dev and sopuli.xyz

I figured that adding some of the major instances should kick-start the federation process and allow it to find other things, but it seems it does not.

Now I’ve been able to subscribe to a couple communities from my instance, memes@lemmy.ml and lemmyshitpost@lemmy.world and of course this Selfhosted community. However, I still cannot seem to add others, for example it still fails to find anything when I try to add memes@sopuli.xyz

Additionally, I can’t seem to see comments from most other users on most of the posts that I do see for some reason. Even if those users are from one of those instances I’ve allowed, I still don’t see them.

Does anybody have any idea why this might be? For example, I’m quite sure I’m not going to see any of the comments on this post even though I posted it, I’ll probably have to come back and check on another account to reply and I really don’t understand why and would like to find a fix.

Thanks.

  • seang96@spgrn.com
    link
    fedilink
    English
    arrow-up
    2
    ·
    1 year ago

    Allowed instances just whitelists them. Federation takes time especially for the large ones to connect to your instance and start working well. You also need to subscribe to remote communities for it to continue updating. There is also currently a timeout that large instances are dropping sending updates (upvotes / down votes / comments) to all connected instances because of them having to update 200k instances for each event. There is a fix in the pipeline to increase the timeout though, so hopefully that’ll be better in the near future.