I ask because I made an account on https://lemmy.zip and the kbin.social magazines there seems to have incomplete contents. e.g. kbinMeta@kbin.social there only have one thread from 25 days ago. RedditMigration@kbin.social there seemed more up-to-date but number of comments are still less than what I can see directly on kbin.social.
Another indication of trouble is that when I try to subscribe to a kbin.social magazine (like kbinMeta@kbin.social) it is stuck on Subscribe Pending seemingly forever and does not transition to subscribed/joined status.
At first I thought it is a configuration problem on lemmy.zip or some recent version Lemmy incompatibility with kbin, but I was able to subscribe to fedia@fedia.io there fine and it seems to be getting updated contents.
So somehow traffic is not flowing smoothly between kbin.social and lemmy.zip.
@ernest, I hate to bother you but are you seeing something like this with other instances?
@1Fuji2Taka3Nasubi Hey, during the server migration, I had to unload the queue (the previous week). I added several addresses to the blocked instances that were either unresponsive or generating errors at that time. Lemmy.zip ended up on that list - probably by mistake. My bad. Now, after the initial verification, most of those instances are still dead, but I will verify it more thoroughly today. Now that I have sorted out the infrastructure issue, a series of federation fixes will follow - perhaps even this week.
Thank you for looking into it. I tried subscribing again and it went through, and several threads just got through to kbinMeta over there, so I'll cautiously say it seems to be fixed now.
Glad to hear federation is being worked on, it does feel it needs to be more robust to live up to the promise that one gets the same experience no matter which instance (local or remote) one view a magazine from. (I'm not being critical, interoperability is always a hard problem no matter which stage of development it is)
Again, thanks for the good work and for bringing kbin to us!
Actually, I have a plan to improve the federation, but at this scale, I need to be super cautious. Just one mistake is enough to harm other instances. That's why I want to prioritize setting up robust testing environments first, so that the every changes in this regard are thoroughly checked and validated.
@ernest Is there just a constant cache of posts to be received that have to process before updating in a magazine here? I'm confused on why some magazines are only behind on syncing with lemmy communities by a few hours, but some are weeks behind on the same instance.