I see a post here: https://lemmy.ca/post/921003. There is only one comment from me, but on lemmy.ml there are more comments and votes. Why is this the case?
yes, we are still federated with lemmy.world.
i haven’t seen anything in the logs, but I haven’t dug too deep yet. i’m going to guess it’s likely an issue with .world themselves. there seems to be a few users reporting issues, i’m guessing because of their high activity/user count:
https://lemmy.world/post/646742
https://lemmy.world/post/574368
https://lemmy.world/post/406956
it could be a general lemmy federation issue too?
Thank you so much for looking into this! You must be swamped with all of us newbies here on lemmy.ca.
If it’s a general Lemmy federation issue. Do you have any guess, if the issue can be resolved at some point? (I’m aware, it might not necessarily be at the top of anybody’s priority list with all the influx happening right now.)
Lemmy isn’t even at a point 2 minor version yet so it will be awhile before it is stable and these kind of kinks worked out, but this kind of thing is going to somewhat the norm I think, by design.
Beehaw defederated Lemmy.world, and they are both have large communities. Although we federate with them both, depending on who/where it was posted the source of truth may result in not always getting consistent updates from both.
I believe the reality in federated space is you will always want to have an account on all the more active instances, using one local as your primary and monitoring things you want to follow on multiple. Luckily the official app makes it pretty easy to do, and maybe there is another app, or one in the works, that will do a client side merge across accounts.
i think there might be a few reasons why this is happening.
from what I understand, all federation requests are stored in a queue, and if the server happens to restart with a bunch of stuff in the queue, then those federated comments/posts/whatever are basically lost. pretty sure they’re working on saving the queue so that that doesn’t happen anymore.
there’s also another issue, linked from that issue i posted above, that has to do with HTTP signature expiration times. i’m just guessing what that means, maybe if it takes more than 10 seconds to federate then the content is rejected?
so it looks like they are actively working on some of these issues.
