First example: Four days ago I posted an article to two different c/collapse communities. If I browse each community on their instance, my posts don’t show up.
Example
Second example: I posted to a community but comments on my post don’t appear unless I go to their instance. When I leave a comment it also doesn’t show up for them.
Example
Can anyone explain what is going on?
In order to catch up with a massive queue backlog a temporary hold on delivery to domains that are being problematic has been setup. Those other domains aren’t responding to kbin.social’s requests properly and so our own queues are getting backed up.
It’ll all be back to normal soon. Once the queues are sensible we can reenable domains as they start becoming responsive to us.
Thank you!
This is a workaround, but… If the original instance you posted to allows global search for non logged-in users, you can click your comment or thread > more > copy url to fediverse and paste that into their search box. It will likely show no results the first time, but click search again and if it was able to pull the comment it should now show up.
I’ve also noticed sometimes my threads (posts) don’t federate out properly, but comments (notes) federate out quickly and those will include pulling the original thread. Sometimes I will add a comment to my original posted thread to bypass whatever issue that is.
In your first example, that second one looks like the lemmy.ml one judging by the icon.
lemmy.ml is blocking kbin users from interacting with it.
As for the first one, is that the sopuli one? They run 0.18.0 according to fedidb, which iirc has general federation issues I think? Not sure what those exactly are and if they’re related, just remember people bringing that up a lot. It’s fixed in 0.18.1 iirc, but that one hasn’t had a stable release yet from what I’ve read.
Edit: Don’t know about the second example, I haven’t heard of any issues with lemmy.world. Did you give it some time? Federation isn’t always very fast.
Edit: lemmy.ml also runs 0.18.0 iirc, so if kbin wasn’t actually involved in the first example (since it sounds like your main account is on lemmy.world), then the 0.18.0 issues still apply to lemmy.ml too.
I think .18.1 reintroduced websockets, which the removal of in .18.0 broke some of the federation. I too have noticed a lot of lag or omission in a lot of feeds. The last I read about it .18.1 was to follow pretty soon, but I also have used kbin primarily for a while so I may be out of the loop for specific Lemmy issues. Once both versions get to 1.0 though we should be in good shape! :)
I just saw someone on a feed saying they’re using .18.1 and asking for any feedback, so it’s starting to get out there in some of the instances.
websockets got reintroduced? sure?
Absolutely not sure lol. But that was my impression of what the discussion was about. At any rate, the change of protocol transmission in .18.0 did cause some problems, so the version under review and testing hopefully will get things back to more functional, however it works.
Apparently adjusting in
nginx_internal.conf
how requests are either proxied to Backend or Frontend depending on the content type helps on0.18.0
.https://github.com/LemmyNet/lemmy-ansible/issues/106#issuecomment-1606222766
Thanks for the explanation! My main is Kbin.social