Strong TPS Lag from a second to other.

I'm having lag problems since 2 days ago. All goes fine (CPU threads: 4, CPU load average: 15%) but from a second to other one of the CPU threads go to 100% and stay there, lowering TPS and making a very strong lag.

I tried with Bukkit "timings" but I get no difference on plugin time usage. I would like to know if there is a way to see where lag comes from (why is this thread getting stucked on 100%). Is there a plugin or something to know which is the reason of that?, maybe a bad chunk?.

Can I know in which chunk is causing this lag?

Thanks for any help. I'm using Bukkit 1.5.2-R0.1

 

jVisualVM can check. What plugins do you use?

 

It look fine... but I haven't got any graphical interface. Using a Linux server.

I'm using:
10/05/2013 - Before this changelog I never had this problem.
* Jobs (Updated v2.11.0)
* FactionChat (Updated v1.51)
* VirtualPack (Updated 2.5.1)
* Orebfuscator (Updated 1.9.3)
* OpenInv (Updated 2.0.2)
* VanishNoPacket (Updated v3.17.3)
* Worldguard (Removed regions from MySQL)
* MobArena (Updated v0.94.4.98)
+ HungerGames v1.3

But like I said before, timings isn't giving any extra time on plugins so I don't know if is a plugin reason or a chunk one.

EDIT by Moderator: merged posts, please use the edit button instead of double posting.

 

I would not recommend using worldguard in SQL mode, it isn't coded well at all.

Orebfuscator can cause this issue. There are ways to have a lag free orebfuscator but I can not discuss them on this forum.

 

I don't use worldguard in SQL mode beacause it gives lag on each modification. I know the way to have the "lag free orebfuscator" but I don't think is a orebfuscator fault. The problem isn't a continuous lag because of CPU being used on a normal way (So lag moments). I have a thread going to 100% and stuck there, whatever the number of players, even if players log-out after the lag the lag still there, the cpu still on 100%.

 

It's a long shot but... I had a similar problem which was caused by a player with a spawner that was spawning hundreds of skeletons and forcing them into one location. i.e., all 1,000 skeletons were at the same block. Kicked the player and killed the skeletons, and the problem would go away.

So maybe you can see if that's happening with your server. Not sure if this would explain the CPU being stuck at 100%; Perhaps if the chunk was still loaded.

 

Orebfuscator is incredibly CPU intensive, so it may be worth testing without it to see If you still lag.

 

Yes, I know Orebfuscator in incredibly CPU intensive, but it goes fine with 30 players... the problem I'm having is with 5, 15 or 30 players, I have it without any apparent reason. Maybe a player login in a problematic chunk (Is what I think, this is why I asked for a way to detect chunks with bigger load times.

 

Have you ever thought someone might just have a redstone clock? Try using the RedstoneClockDetector plugin or something similar.

 

Yes, I though on it, but I didn't think clocks detectors existed :S. I'll try one out.