Nutomic

This seems like a bug in Lemmy so youre definitely right to report it here. Doesnt matter if you observe it from another platform. It looks like all requests go...

The bug also seems to happen for instances that are running only a single Lemmy backend, eg [mander.xyz](https://lemmy.ca/comment/7467127).

> It might be relevant since in general every actor has their own inbox, and the merging of these inboxes into a single site-wide inbox might not work correctly depending...

To fix this we have to narrow down what exactly is causing the duplicate activities. My understanding is that the bug doesnt happen with the default ansible configuration, and affected...

I added some extra logging in https://github.com/LemmyNet/lemmy/pull/4726 to figure out where the problem is coming from. Once that is merged and released, run Lemmy with `RUST_LOG="warn,lemmy_server=debug` or `RUST_LOG="warn,lemmy_server=trace` (preferably with...

@MrKaplan-lw Yes that definitely looks like there is more than one worker active. So far its not clear where exactly in the code this is caused, but maybe my changes...

I think that is normal as Lemmy processes a batch of outgoing activities, then goes to sleep for 30s, new activities get generated and then sent out after the sleep...

That particular logic was removed [here](https://github.com/LemmyNet/lemmy/pull/4733/files#r1609799891). So if were lucky the problem is already fixed in 0.19.4-rc.3.

It should be enough to throw an error [here](https://github.com/LemmyNet/lemmy/blob/main/crates/apub/src/objects/instance.rs#L208) if the site has a content warning. Then it wont be possible to fetch any communities from nsfw instances.

This one has a lot of conflicts. Should also rename migration files to the current date.