Problem General LAG and Server stops every AM

[netmonkey]

•What OS are you running?

Linux ftb 2.6.32-5-amd64 #1 SMP Sun Sep 23 10:07:46 UTC 2012 x86_64 GNU/Linux

java version "1.6.0_26"
Java(TM) SE Runtime Environment (build 1.6.0_26-b03)
Java HotSpot(TM) 64-Bit Server VM (build 20.1-b02, mixed mode)


•Are you using minecraft hosting provider or a dedicated/local server not designed for minecraft?
The server is hosted but we control the whole OS and software / configuration.


•What version of FTB are you using? (Dont just say latest tell us the version)
I have the server updated to the recommended version as of 1am on the 29th Jan. It is 1.4.7, Mindcrack v8.1

•Did you add any mods to the server pack?
Yes the following mods have been added:-

Within Mods

AdvancedPowerManagement-1.1.55-IC2_1.112​

DamageIndicators[1.4.6]2.3.7​

Within core mods

[1.4.6]TreeCapitator.Forge.1.4.6.r03.Uni.CoreMod​

•What are your server specs?
CPU: Intel® Core™ i7-3930K Hexacore incl. Hyper-Threading Technology
RAM: 64 GB DDR3 RAM
HDD: 2 x 3 TB SATA 6 Gb/s HDD 7200 rpm (Software-RAID 1)

The FTB server is running in a Virtual server with 8gig allocated with plenty of disk space. We are not running a Ram disk. The server tends not to peak over 4gigs of ram usage but all 8 cores are used albeit one always seems to run flat out.

•A good detailed paragraph about the problem

The server accommodates 5-15 people daily with a ramp up of usage towards midnight GMT. We notice a creap of LAG irrespective of numbers of users and this can happen any time of the day. Normally a reboot of the server and minecraft instance restart resolves this. Sometimes it makes no difference.
Also by about 8am GMT we normally find the Mindcraft server is down as most are sleep and have not spotted it. A simple restart resolves the issue.

The world was started on New Years Day! Is now about 1.2 gig big.

My questions are:-

• Why is the server rebooting all the time? It’s happened 3 mornings now on the trot.
• What can we do about the general lag? Surely the server specs have enough grunt.
• How can I optimise or improve the general upkeep. I.e. Am I running the correct version of Java? Are the Mindcraft java parameters correct etc.

Server.log paste >> http://goo.gl/YJiCo

@ 3.48am the server crashed.

A file called hs_err_pid3882.log mentions a JAVA issue.
http://pastie.org/5944283

An example of a htop running

The server is ftb.bigclan.co.uk and to join you will need the 2 following mods installed >>
https://mega.co.nz/#!Q0pBmLBJ!QNr2EE1xOweonosX3xpwRQ2vBpYIN9ml6FjM7urDN8Y

 

[Dekar]

I don't know the exact reason for your issues, however you can try two things.

First, update to java 7 and out of java 6. There are arguments for both versions, but 7 has been more stable for our server.
Secondly, you can try launching the server with less java tags and just specify the min/max ram and not the other items. If this fixes your issue try adding in select items into the start up script to see which ones are causing the issue.

 

[netmonkey]

ok thanks for the feedback. I was looking to install Java 7 prior to setting up FTB on my Debian server. I got the impression that Debians source lists were not setup for Java7 yet? I also have a test server which is a replica of the Live server which I use for patching and testing. I may consider another distro if I cannot get Java7 installed.

Apart from that is there a case to go with RAMDISK to improve I/o with the Disks?

 

[Dekar]

I am using Debian 6 on my server and have java 7.10 I think the last time I checked. Might have updated to .11 can't remember. I never saw much issue with I/O unless creating new chunks and such is giving you a lot of lag. Because that is the only time that really creates that much data to be written tot he disks.

 

[netmonkey]

Ill aim to get JAVA 7 installed and take it from there.

Any recommendations on the parameters to set against the Jar.

I do wonder though if the lag and the fact the Minceaft instance is prone to stop is due to the way we play the game. A lot of the players have at least 2 Chunk loaders running. Giant Quarrys have been dug out and there is plenty of Cobblestone and Pneumatic piping everywhere. Alot of the guys have spread about the world also.

 

[Dekar]

I am using the default java commands from the launcher basically. I am on a 2g vps, so I moved the memory up a bit. The only real error I get is a memory issue, which shouldn't be an issue with yours since you have a lot more ram to play with.

The parameters you are using should make java run faster, but it could be one of them that is causing java to crash too, which is why I recommended a processes of elimination for them.

Our current world is over 1gb in size with 20+ players, but usually under 10 online at a time. There are several chunk loaders that I am aware of and we made a mystcraft world specifically for quarries, which have been used extensively. I did run into a few issues recently with those, is that having a 64x64 quarry filled with running water seems to effect the server. We had about 5 quarries all basically filled with water and I was getting random minecraft crashes and people timing out. After clearing the water (which caused a bunch of crashes in of itself) I have noticed a decent drop in issues with the server and lower cpu/mem usage. The other thing is that I saw posted, quarries that reach bedrock and are left idle continue to try to run which uses cpu power in addition to keeping chunks loaded.

So, I made two new rules for our server and that is, no water in quarries, if you hit lava take care of it at that point and remove the water once the lava is gone. And once a quarry is finished to bedrock take it down so that it is not loading chunks that don't need to be.