Is anybody else running into this problem?
My federated posts are appearing in my timeline quite well, however I’ve noticed that significant portions of the comment section are not being carried over.
For example, I’ll see a post on my instance with 13 comments. I’ll click on the fediverse icon to take me to the original source and that has 43 comments. Where did those 30 comments go?
The worst part is, I can’t seem to figure out a way to “force sync” or refresh the page to make sure all the comments are carried over. Because of this, I feel like I’m only getting part of the conversation and I’m unable to reply to comments that don’t appear on my instance.
Does anybody have any solutions to this problem? It’s literally my one huge issue I have with my own instance.
I have the same issue. I use multiple accounts across different servers, and get a different subset of comments across each one. Not sure the solution right now.
Thanks for pointing this out, I am just now realizing I am also missing comments. For example this post, I have 7 comments, but the actual post has 10.
I’m glad I’m not the only one. I am commenting on this post from my own instance and it says it has 0 replies which I know is wrong. I did only federate with Selfhosted just a few hours ago. Who know how long it will take this post to get back to lemmy.world.
Yep, it’s because of the volume of comments/posts going through the big instances right now, and how many people are subscribed outside of those instances. It’s too much for the server to keep up. I noticed over the last day or two my feed has mostly been Beehaw.org stuff because they don’t seem to be having as hard of a time. It sure is looking like we need more separation like startrek.website, and maybe it was a bad idea to allow communities to just be made by anyone on lemmy.world and sh.itjust.works. Would’ve made more sense for all the users to be on these instances, but the communities hosted elsewhere.
Not sure if this is the same problem, but it has been noticed that passing messages between federated instances has been hampered by the way the “worker threads” are currently been done. The big instances are essentially getting overloaded with outbound federation work. Solutions for that are in the works. issue/3230
Hopefully it’ll mostly resolve given time. Just like when there was a mass exodus to mastodon, the major thrediverse servers are getting overloaded. This creates delays and it seems even lost messages. From experience I’ve had messages from fedia.io take anywhere from 20-120 minutes to show up on my mastodon. As admins beef up their servers, but also as excitement tones down and there is less enthusiastic usage, hopefully the problem will resolve for the most part.
I’m not sure what causes messages to be lost forever vs. just delayed. Maybe Kbin or Lemmy aren’t quite as robust in that? When mastodon was melting down, messages could take 8+ hours to get through, but they seemed to get through. I think your idea about a resync idea is a great one.
Yeah I started on a dinky little server and now it’s massive… mostly seems to be a memleak or something going on though that’s causing OOM and a massive CPU spike. Anyone know anything? 4GB of RAM getting eaten for breakfast… just had to up it to 8
I’m curious. How many users in that instance? I assume that an instance for a few people don’t eat that amount of ram.