-- Snip --
@CyphurTheFox Interesting results, thanks for taking the time do look into that
Have a llama in thanks
1. Your estimation of TPS is correct. I did a timing breakdown over the past hour, and the average TPS was 2.2. Yesterday when I checked it was around the same, but it's good to know that you've confirmed that it's like that all the time, and not just in spikes.
2. Lag being present regardless of location is interesting, as (for the most part), it eliminates structures or entities as the cause, unless they're in the spawn chunks. Based on the timings log I collected, it seems sounds and pistons are taking longer than they should, but I'd be willing to bet that's just a side effect of everything else lagging
3. I never manage to catch the restart, so I was unable to test this myself, so that's good to know. I wasn't sure if the restart fixed it and it gradually got worse over the news few hours, or if it was an instant effect
About the map:
- I've got the plugin set to only render the map when the TPS is above 15. If it ever drops below that, it'll pause the map until it stabilises again. Similarly, it should pause the map rendering if there's more than a few people online, to avoid lag then too
- From the timings report I got, dynmap is taking up 1.51% of the tick, which while that's high enough to give me a yellow warning, it's nowhere near high enough to be the source of the lag
- The
saverestorepending flag is set in our config, so it should be keeping its progress over restarts
- From what I can tell though, the queue is empty, so even if the map were a problem, it shouldn't be rendering anything right now
Based on the timings I took, the things that are taking up the most time are
doTick and
doTickPending, which are taking up 825% and 768% of the tick time each, respectively. The
doTickPending I'm pretty sure is only taking up so much time because it's being postponed by
doTick taking longer than it should. I'm trying to see what I can do about this, but there's not too much info out there that I can find. We had this issue about 2 years ago, which our host tried to resolve by moving some stuff around and seeing what could be done, but from what I remember it just randomly fixed itself. I'll try and keep you updated with any progress I make on fixing it.
If anyone notices anything else that could be useful, please let me know
Thanks, and sorry again for the lag