What's happening with our All/Hot? (i.postimg.cc)

submitted 1 year ago* (last edited 1 year ago) by oxideSeven@sh.itjust.works to c/main@sh.itjust.works

17 comments fedilink hide all child comments

I tested this on multiple browsers and systems. Our All/Hot has been like this for 2 days. If I leave it long enough, multiple minutes it will sometimes load a bunch more and get stuck in weird update cascade that will use up all my RAM.

None of the other instances I tried have this issue. It seems to happen whether I'm logged in or not.

top 14 comments

sorted by: hot top controversial new old

[-] simple@kbin.social 14 points 1 year ago

It's a known issue with Lemmy, it should be fixed by next release.

[-] Barbarian@sh.itjust.works 10 points 1 year ago

0.18 is gonna fix so many issues. Glitchy feeds when a new batch of federated data comes in, glitchy vote numbers on posts, performance improvements, post ranking on feeds, etc.

[-] oxideSeven@sh.itjust.works 4 points 1 year ago* (last edited 1 year ago)

But I only see it happen here. Are there other instances with the same issue? Or a github post about it, do you know?

[-] Barbarian@sh.itjust.works 3 points 1 year ago* (last edited 1 year ago)

Does it only happen here? Paging @TheDude@sh.itjust.works and/or @manifex@sh.itjust.works. Is there something wrong with our hot_rank column in the database? Is the update not triggering?

EDIT: Jumped on the Matrix chat to ask

[-] oxideSeven@sh.itjust.works 3 points 1 year ago* (last edited 1 year ago)

As far as I can tell, yes. Only here. The screenshot shows FMHY updating, and I checked a few more before hand.

[-] Barbarian@sh.itjust.works 9 points 1 year ago

Got an answer from TheDude on matrix chat. That is a known issue fixed in 0.18, but some instances are using the untested patch to fix it until then. He doesn't want to roll out untested code with potential side-effects, so we're stuck with the issue until it comes down the pipe.