could random high tick times be caused by using a shared host?
my server has perfect tps and perfect mspt averages. its just that every 2000-10000 ticks or so, a tick that is way longer than usual is processed ( from around 1-10mspt to 160mspt ). this happens just for one tick and returns to normal. i dont think i have any plugins that could be causing this (all my plugins from spark profiler are below 3% in tick occupation). the only reason i can think of that causes this is shared hosting with other nodes using high resources
35 Replies
Thanks for asking your question!
Make sure to provide as much helpful information as possible such as logs/what you tried and what your exact issue is
Make sure to mark solved when issue is solved!!!
/close
!close
!solved
!answered
Requested by kylosoot#0
bump
how much RAM does the server have allocated - it sounds like a garbage cleaner issue
actually a spark profile should be able to diagnose it if it's GC
8gb
this happens even with only one person online
try recordin and sending a Spark profile
okay
seems like your right
do those coincide with the spikes?
yeah
hmm
if the GC is going off too often, it might be some plugin eating too much RAM
what plugins are you running?
[08:57:13 INFO]: Paper Plugins: (2):
[08:57:13 INFO]: - FancyHolograms, HuskClaims
[08:57:13 INFO]: Bukkit Plugins: (64):
[08:57:13 INFO]: - AggressiveAnimals, ajLeaderboards, ajParkour, AntiPopup, Back, Chunky, Citizens, CommandPanels, CommandWhitelist
[08:57:13 INFO]: ConditionalEvents, CoreProtect, Craftorithm, CustomCommands, CustomDrops, CustomJoinMessages, DiscordSRV, DiscordSRVUtils, EasyCommandBlocker, EnchantsSquared
[08:57:13 INFO]: ExcellentCrates, ExtraContexts, floodgate, GCore, Geyser-Spigot, GeyserUpdater, Hurricane, InstantCommand, InventoryRollbackPlus, InvSeePlusPlus
[08:57:13 INFO]: InvSeePlusPlus_Clear, InvSeePlusPlus_Give, ItemEdit, ItemTag, LagFixer, LPC, LuckPerms, LuckyTPA, Multiverse-Core, Multiverse-Inventories
[08:57:13 INFO]: nightcore, Oneblock, Oraxen, PerWorldPlugins, PistonMOTD, PlaceholderAPI, PlayerKits2, ProtocolLib, PurpurExtras, ServerLinksZ
[08:57:13 INFO]: ServerVariables, Shopkeepers, *Sudo, TAB, Tweakin, UltimateHomes, Vault, VeinMiner, ViaVersion
[08:57:13 INFO]: Vulcan, WorldEdit, WorldGuard, WorldGuardExtraFlags
actually if 103.02ms is a 106% increase from average then your average is like 50ms (anything higher gets below 20tps)
i'm not too experienced with specific plugins but i've heard Citizens is not good for lag - a full Spark profile would help see what's lagging
i dont know why it says that but my actual mspt is like below 15
im running the spark profiler atm
hmm
when you end the profiler, send it
how long shouuld i run it for?
a few minutes, maybe
if you want to try to post a spark link with a big lag spike, do this
https://spark.lucko.me/docs/guides/Finding-lag-spikes#step-2-use-spark-profiler-with---only-ticks-over-to-find-the-cause
Finding the cause of lag spikes | spark docs
Lag spikes occur when a small number of ticks (or sometimes just one tick) takes a long time to execute.
I'd say use the tickmonitor at the same time and include like 5 or 10 lag spikes if possible
@If you want we can transfer your server to my vps with dedicated rescources for a quick test, i'm running windows server however so unless it's in a docker container we may run into some issues.
(i have WSL installed if it becomes neccessary)
Yea no
Don't do that.
Don't transfer your server to random guys vps
hes already done it once in https://discord.com/channels/348681414260293634/1310913654463401984
Really sus.
can agree
Sounds like a provider lowkey trying to advertise.
or just want to inspect your files
That too and probably worse.
news flash, if you are on someone else's vps, they can see everything
doesnt have enough data, those lag ticks (over 150) only happen around once every 1-2m so its really hard for the profiler to get enough data
then just let it run overnight, or multiple hours
Not advertise since I'm doing it for free for smaller servers, I was just trying to get some experience in this. The mods have already talked to me about it and we sorted it out.
I won't be offering that anymore, cause I do genuinely see yalls concern.
But yall can talk to shing and his friends if yall are worried bout em
https://spark.lucko.me/hOUdzVtAV1 bit late cus didnt have access to my server
spark
spark is a performance profiler for Minecraft clients, servers, and proxies.
An error has occurred!
This was most likely an error on our end. Please report this at the github
everything is normal except the max mspt. i was wondering how to lower it because every 1-2 minutes it spikes up to 100-250ms for one tick
that one wasn't started with
--only-ticks-over :(
it also doesn't show that a single tick went over 100msptoh yeah my bad
it didnt say it in the report but it did say it in the
tick monitor ingame