EDIT: Issue “fixed” itself after a few days, see comments.
I’m aware of the ongoing issues with federation being a bit patchy, and have got used to manually searching for posts / comments to pull them in and sync them. No problems there, usually.
But the last couple of days I’ve seen two threads in particular that seem to work everywhere else, but never made it to .world and can’t be pulled in manually either.
The first was https://lemmy.world/post/1005008 (throws an error, but visible at https://lemmy.blahaj.zone/post/600356) which also posted multiple times to the community it’s in so I thought the issue was maybe something to do with that.
The second is https://lemm.ee/post/703295, which did not have any multiple-posting issue as far as I can tell.
For both of these threads I’ve tried manually searching for the post AND for comments. The former allows me to search up a reply, but then gives a couldnt_find_post. error when I try to click through. The latter doesn’t find any replies at all.
Just thought this might be an interesting data point as it seems to be different to the typical problems we’ve been having. If anyone has any idea what’s broken with these two posts I’d love an explanation!
Ok this is an ongoing issue, post at https://reddthat.com/post/354099 has the same problem in that it’s federated elsewhere, but not here, and I’m also unable to manually sync it. Note: This one was also accidentally posted multiple times to the same community, which makes 2 out of 3 of the problematic posts.
Not sure if it’s a Github-level issue since I’ve only seen it happening to us on .world specifically, I’ll have a look around the issues list and see if anyone’s reported similar from elsewhere.
Day 3 update: seems to have fixed itself? As of this morning I’m now able to manually pull in all of the mentioned posts.
Perhaps it was an incompatibility between the different server versions or something, as most involved have upgraded to 0.18.1 since yesterday. I did raise a Github issue which has no answers either, guess this update is just for posterity in case the same thing happens in future and anyone is looking for info on what “solved” it in the end.