Lot's of micro-laginess in various situations. But I cannot really quantify if it worse or better than before, NS2 was always a "game of thousands little lags".
Just updated right now and my FPS dropped to unplayable 10-20 . It never was good honestly but now I'm worried this will stay until a performace fix patch Probably the new collision models.
AurOn2COOKIES! FREEDOM, AND BISCUITS!AustraliaJoin Date: 2012-01-13Member: 140224Members, Forum Moderators, NS2 Playtester, Forum staff
Doesn't justify the use of the spam flag. Point out their mistake and move along. or at least make it painstakingly clear what you are talking about. Thread is called "b61 performance", post could easily be taken as a performance post excepting the fact you snuck the word server in before performance once.
AANyway server performance has definately gone bonkers. Deploying hotfix... soon?
Also redplug at round end is back, plus occasional lagspikes.
Other servers are affected too as far as I can tell from the server list.
ditto ☺
Servers that were stable and had capaticies above their slot number are now not running at 30 tick rate anymore. In some cases there are very severe drops.
Also occasional server lag spikes and occasional "red plug" at round is back from what I can tell.
My server logs now have massive amounts of sound errors running combat mod, sound for co has not been recompiled in several builds and was no issue until now.
Okey this is weird but, i've gotten a fps increase in this build. +20-30 ballpark. It's been weird holding 200fps on some readyrooms before players load in .
I also noticed a drop in server performance, those pubs i've played before and even competitive servers for pcw i do suffer of small "micro-stuttering"... What's been changed in 261 on the server side that made 260 run better?
Ups my mistake, however the title is not precise. Who would think on first glance that this is about server, when its presented in "Getting Started" and noone sees its in "Server Discussion" subforum?
Marking polite and on-topic responses as "spam" is IMO over the top. So I reciprocated with an abuse flag to brighten up your day.
How exactly are you going to determine whether a rubber-banding instance is client or server issue?
It's not on topic at all. It's completly off-topic. Mods should remove those posts and put them in new thread if people want to talk about client performance, but this is detrailing a thread with the first posts, thus spamming.
By looking at server tickrate and CPU usage monitoring (htop, top etc) per server_linux32 proccess (or the windows equivalents).
Ups my mistake, however the title is not precise. Who would think on first glance that this is about server, when its presented in "Getting Started" and noone sees its in "Server Discussion" subforum?
Anyone else noticing a considerable drop in server performance in B261?
(also, still need better multi-core support on servers and generally better performance--)
But at least this proves a point which is that the forum community doesn't read posts carefully or even at all. Hell, if even two lines is enough to not have them read, why even post long posts?
dePARAJoin Date: 2011-04-29Member: 96321Members, Squad Five Blue
I didnt notice any issues on my server.
Same ammount of RAM and CPU used like the last builds.
One server can go up to 28% CPU usage and the rookie wave is filling up the servers.
If you run 5 full servers on an quadcore you lose cpu power for an single server for sure.
Try to reduce the serverammount, im sure this help.
I have noticed rubber-banding and severe lag/loss on two servers I frequently play on in the in the last couple of rounds.
One of the servers was restarted by the admin, about 15 min. into the next round players started to notice the lags again, and they got worse as the game went on.
Both servers run 16 slots, shine admin and (in one case) NS2State for Shine as mods.
SamusDroidColoradoJoin Date: 2013-05-13Member: 185219Members, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Gold, Subnautica Playtester, NS2 Community Developer, Pistachionauts
edited December 2013
We ran with shine + TGNS and then Vanilla, experienced same symptoms, jitteryness. Apparently PT's believe it has been happening since before it collision was added, which is really the only possible thing it could be.
@Solaris This also happens at vanilla servers and i'm pretty sure it's not related to one of the given mods.
I'm not suggesting it is (looks like ye olde memory leak to me tbh), just giving extra information that - from my limited sample size - seemed like it could be useful, if only to falsify a thesis.
Yes, it seems like the server has memory leaks. Seeing the memory usage it has climbed close to 2GB for my servers again and I just had one crashed that seems to be due to memory allocation (on mapchange). Which also might be the cause of lagspikes
#0 0xf6be7834 in lj_BC_UGET ()
from /home/gameservers/common/games/ns2/liblua51.so
(gdb) bt
#0 0xf6be7834 in lj_BC_UGET ()
from /home/gameservers/common/games/ns2/liblua51.so
#1 0xf6b86e7d in lua_pcall (L=0xdeab300c, nargs=1, nresults=0, errfunc=4)
at ../Source/LuaJIT/src/lj_api.c:1111
#2 0x080652fa in M4::Script_Call (L=0xdeab300c, nargs=1, nresults=0)
at ../Source/Spark_Server/Script.cpp:1672
#3 0x0809915e in Callback (this=0xed3d3df0)
at ../Source/Spark_Server/World.cpp:182
#4 M4::World::HTTPRequest::Callback (this=0xed3d3df0)
at ../Source/Spark_Server/World.cpp:161
#5 0x080a0c48 in M4::AsyncTaskList::InvokeCallbacks (this=0xed2c0df4)
at ../Source/Spark_Server/AsyncTaskList.cpp:55
#6 0x08074d8a in M4::ServerGame::Update (this=0xed2c0ca8,
deltaTime=0.0030378869996638969, waitForPackets=true)
at ../Source/Spark_Server/ServerGame.cpp:363
#7 0x080719da in M4::ServerEngine::Update (this=0xf6b721ac,
deltaTime=0.0030378869996638969)
at ../Source/Spark_Server/ServerEngine.cpp:70
#8 0x08061f04 in M4::BaseEngine::RunFrame (this=0xf6b721ac)
at ../Source/Spark_Server/BaseEngine.cpp:196
#9 0x0805db28 in M4::Server::Run (this=0xf6b72170, settings=...,
commandLine=..., console=false, webServer=true,
modDirectory=0x814fb3c "/home/gameservers/common/games/ns2/mods/")
---Type <return> to continue, or q <return> to quit---
at ../Source/Server/linux/Server.cpp:166
#10 0x0805c06e in Run (parser=...) at ../Source/Server/Main.cpp:166
#11 0x0805af06 in main (
argc=<error reading variable: Cannot access memory at address 0x6>,
argv=<error reading variable: Cannot access memory at address 0xa>)
at ../Source/Server/Main.cpp:224
I've also seen the entcount climb close to 1000 when servers were lagging, so there might be some ingame issue with entities not being removed properly as well? O
Comments
Also redplug at round end is back, plus occasional lagspikes.
Other servers are affected too as far as I can tell from the server list.
He asked about server preformace, not client preformace.
AANyway server performance has definately gone bonkers. Deploying hotfix... soon?
ditto ☺
Servers that were stable and had capaticies above their slot number are now not running at 30 tick rate anymore. In some cases there are very severe drops.
Also occasional server lag spikes and occasional "red plug" at round is back from what I can tell.
My server logs now have massive amounts of sound errors running combat mod, sound for co has not been recompiled in several builds and was no issue until now.
How exactly are you going to determine whether a rubber-banding instance is client or server issue?
Uh, because rubber-banding is only related to server performance.
It's not on topic at all. It's completly off-topic. Mods should remove those posts and put them in new thread if people want to talk about client performance, but this is detrailing a thread with the first posts, thus spamming.
By looking at server tickrate and CPU usage monitoring (htop, top etc) per server_linux32 proccess (or the windows equivalents).
a) It's right there? and
b) Read the god damn post(s)
But at least this proves a point which is that the forum community doesn't read posts carefully or even at all. Hell, if even two lines is enough to not have them read, why even post long posts?
Same ammount of RAM and CPU used like the last builds.
One server can go up to 28% CPU usage and the rookie wave is filling up the servers.
If you run 5 full servers on an quadcore you lose cpu power for an single server for sure.
Try to reduce the serverammount, im sure this help.
I left it up right after 261 went live and it had some pretty bad tick rate issues even at 16 players.
One of the servers was restarted by the admin, about 15 min. into the next round players started to notice the lags again, and they got worse as the game went on.
Both servers run 16 slots, shine admin and (in one case) NS2State for Shine as mods.
I've also seen the entcount climb close to 1000 when servers were lagging, so there might be some ingame issue with entities not being removed properly as well? O