What's up with "The magazine from the federated server may be incomplete. Browse more on the original instance."? Is this working as intended or is it a bug?
What's up with "The magazine from the federated server may be incomplete. Browse more on the original instance."? Is this working as intended or is it a bug?
I've been seeing this for quite a bit, and thought it'd resolve itself once CloudFlare was taken off, but I'm still seeing it on many external communities from e.g. Lemmy. Not all posts or comments are visible from Kbin. Any idea on what's going on?
Edit, clarification: when I visit e.g., https://kbin.social/m/[email protected] the above mentioned sentence shows up, and I'm not able to see threads created even 13 hrs ago, like this one on the original instance https://beehaw.org/post/927935?scrollToComments=true. New content doesn't seem to be pushed to this instance at kbin.social. This is the issue I'm referring to.
Edit 2: This is probably the answer: https://kbin.social/m/kbinMeta/t/131241/What-s-up-with-The-magazine-from-the-federated-server-may#entry-comment-517045
You're viewing a single thread.
I'd like to know exactly how this works as well. I've seen people say there's no real downside to joining a smaller kbin instance as you can still get to other magazines in different instances, but this seems like it would be a drawback if you can't browse old threads in a remote magazine. Or would it eventually pull all the threads in? Would be nice to know what the expected behavior is.
I've noticed that sometimes new threads from federated magazines sometimes don't show up for me in Kbin, as well as older ones. I think there may be some sort of syncing issue that causes that, perhaps Kbin refreshes data from other instances at set intervals or something.
Yes, this is the issue I'm referring to!
One thing that also plays a role is the size of the queues on the different instances. Kbin.social may get things pushed to it, but with queue sizes reaching 500 thousand or more, it takes time to process them.
I think it's actual data loss, not just lengthy queue times. I see new threads and comments appear but for some of the older threads, even days after they were active, I see missing comments if I compare it between kbin.social and the original instance truth.