Lemmy starts out pretty chill, but the longer it is running, the more CPU it’s using. Until I restart it again, and then the process starts over again. It’s just /app/lemmy that’s eating the cpu cycles. Looks like there are some threads that keep requeueing themselves, until eventually that’s all it’s doing.
Does anybody have any clues or pointers about this?
I’m running 0.17.4, haven’t made the jump yet to 0.18.
What is your server doing?
Lots of local users logging in and reading content? Federation outbound sending of new content from your server to other servers? Federation inbound?
On my instance, i have a lot of federation incoming messages, subscribed to every large community I could find. And I’m getting posts, comments, votes coming in constantly and the CPU usage for lemmy_server is pretty high/constant. The digital signature and boilerplate federation text is actually high overhead for a single upvote to be processed, etc.
But CPU usage goes go down if there are no messages for a minute, but 20+ servers sending to my server constantly, it isn’t idle for long.
I’ve been running the 0.18 docker build since it was released. Just took a look and it’s CPU usage has remained <3% on load since the release. Admittedly my instance has only a handful of users, but it might be worth upgrading to 0.18?
Yeah, my instance doesn’t have a lot of users, but it does have a lot of posts (few thousand per day). It might be all the updates it’s sending out.
I guess I’ll just have to give 0.18 a try then.
I can only speak to my instance but I’m not seeing anywhere near that amount of utilization. I am on 0.18.1.
I restart my lemmy container once an hour.I saw the lemmy.world admin say that’s what he was doing back whenever the new sort algos came out. I still have mine doing that. @ruud@lemmy.world
Sorry, I cannot find the comment now , one of the Android app developers responded to a ‘my phone is getting hot’ comment that caching was the problem. Perhaps this is related?
