Log spamming
gkchp opened this issue · 4 comments
server keep spamming logs like this on startup, no player joined nor any frames placed
latest.log
[19Jul2024 01:57:48.004] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42689115ms or 853782 ticks
[19Jul2024 01:57:49.166] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690277ms or 853805 ticks
[19Jul2024 01:57:49.340] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690451ms or 853809 ticks
[19Jul2024 01:57:49.389] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690499ms or 853809 ticks
[19Jul2024 01:57:49.422] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690532ms or 853810 ticks
[19Jul2024 01:57:49.452] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690562ms or 853811 ticks
[19Jul2024 01:57:49.469] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690579ms or 853811 ticks
[19Jul2024 01:57:49.482] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690592ms or 853811 ticks
[19Jul2024 01:57:49.492] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690602ms or 853812 ticks
[19Jul2024 01:57:49.501] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 42690611ms or 853812 ticks
I think this bug is still present but in a smaller way or a slight regression happened on the move to 2.1.x,
I'm running 2.1.3 (Forge 1.20.1) and get the same
[00:20:51] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 2020ms or 40 ticks
[01:05:51] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 2051ms or 41 ticks
[01:50:51] [Server thread/WARN] [waterframes/]: Server seems overloading, jumping 2086ms or 41 ticks
But the server is empty, no one on it.
Spark profile during the timeframe shows
MSPT
0.264 min
0.322 med
0.393 95%ile
225 max
And TPS being 20.00 for 5/10/15m
during the time period waterframes complained about server seemingly overloading.
I just noticed as I was typing this that 2.1.4 / 2.1.5 just came out in the last few hours, does any of the changes contained in that address how waterframes sees server overloading?
I didn't thinked about but internally WF ponders about any delayed millisecond as a lag spike and stores the time, stacking it at the end.
In short, the reason of the occasional overloading log is by the milimétric precision on lag detection.
In any case, I added a config in serverconfig to turn off the lagTickCorrection. Just turn it on when your server is laggy (19 TPS is laggy enough)