[Minecraft] FTB Announcement - Chunkloaders disabled.

Well, despite the mythbusting thread, somebody's gone and herped the derp and installed a chunkloader on some sort of machine setup which makes the server crash on startup.

Until further notice, chunkloaders have been disabled.

 

No.

 

That is unusual, given that portal and chickenchunks are completely separate mods, there's no reason for those things to be related, and if they are, I don't understand how. :/ I'll have to putz around with it when I'm next online.

Even if we could enable loaders on a per-player basis easily, it's only a matter of time before it causes more problems → Even if everyone keeps it within limits, our server is reasonably popular now and as we get more and more people, more and more chunk loaders means more and more shit the server has to keep track of at the same time, eating more memory, CPU, and other resources, which equates to more and more lag. People pulling stuff like this only makes us reach that point faster.

Looking back at my OP though, I should probably clarify that loaders are only disabled when you are NOT connected to the server - that is, your loaders should become active once you connect, so if you use nether lava as a power source or some such it shouldn't affect your machines while you're online.

It sucks, it's not fair, and it's a PITA, but unfortunately I don't see any alternative. I'm open to suggestions though.

 

They can't be distinguished in the configs. It's both, or neither. There could be a way to disable the recipe, but spot loaders wouldn't have prevented the problem this morning that was the reason they're currently disabled. (see below).

It's not that simple. The server already has nearly ALL the 12 GB RAM on the box allocated to it.CPU usage is the problem -> Minecraft only uses a single thread, and can't make use of multiple cores. That's your bottleneck, the issue isn't a lack of resources, it's that minecraft is poorly written and can't actually USE them. So 3/4 cores are sitting idle while minecraft pegs the last one at 100%.

You also seem to misunderstand the loader problem. The last major bout of lag was caused by leaking worlds. Placing chunkloaders SOLVED that problem by preventing worlds from ever getting unloaded and leaking.
Loaders also weren't disabled today because of lag, they were disabled because someone had shit going on in a loaded area that was crashing the server on startup. will look at re-enabling them once we locate the problem and fix it.

 

I'd say that affects more than one person, no?

 

No, you're not understanding how this affects EVERYONE.
The SERVER was crashing on startup. that means NOBODY can connect and play.

 

The community currently runs 2 full hardware servers. One server runs all of the TF2/Source Engine servers, teamspeak, and mumble. That's 14 different services running on an 8-core server with 12GB RAM. It's not *pegged*, but there certainly isn't room for anything else. The other server is our webserver. It runs the web service that runs our site, the database server, and the minecraft server. This isn't pegged, but we've had spikes in web traffic here and there that pop the load up.

What you don't understand is that we know how to do our jobs. We are in the positions we are in for a reason. Questioning VintagePC on how things work is not helping anyone. When we shut down the Tekkit server, I immediately devoted all of the RAM that was for Tekkit over to FTB. This only helps to a certain extent because of the following reason:

A: Minecraft is written in Java
B: Java is a programming language that makes itself OS-independent through the magic of making a virtual server with its own mini-operating system within your operating system (Inception anyone?)
3: The virtual machine they run is only capable of utilizing a single CPU thread.
C: This makes Minecraft a bloated, buggy, memory-leaky piece of shit that eats the fuck out of resources, except for CPU because it's too mentally handicapped to know what to do with all of the CPU you offer to it.
F: Take that bloated, buggy, memory-leaky piece of shit, and add 45-ish different mods that were written by different people that don't give a crap about how their mod plays with everyone else's mods.

THIS is what we're trying to maintain on our server. This is what is currently using an entire 4GB of RAM at any given moment, and is trying to use 140%+ of a single CPU core.

NOW. Take all of this information, and consider this. This server has 2x 4 core CPUs with hyperthreading. If we were capable of giving it more resources, we would. It will use 100% of the RAM I give it regardless of the amount. I've given it 6 and even up to 8GB RAM, and it didn't increase performance, so I put it back to 4GB, which is where the performance leveled out. It's using every bit of resources that it knows how to use. If we could make it run better, we would. If we could dump more resources at it, it would have been done already. We *HAVE* the resources, it doesn't help anything.



And no, I have no idea wtf you mean when you say it only effects 1 person. When that person logs into the server, it is not their minecraft client that crashes. The console on our server, the part that everyone's client connects to, throws a big ole capital-letter "JAVA FATAL EXCEPTION ERROR: I'M A BIG PIECE OF SHIT BECAUSE JAVA" and dies. It passes on. It ceases to be. It rings down the curtain and goes to join the choir eternal. It is an ex-server. Then one of us (the server admins) has to log in and restart it so you can connect again and play.

 

I don't know where you're getting your information. I also don't know why you'd choose to believe your "source" rather than what's coming straight from the staff members involved in running the server, but I'm going to explain things again, with supporting evidence and correct a few misconceptions below.

Let's get to the facts:

1. The server was NEVER placed on an auto-rebooter. It was SUGGESTED in a thread as a fix for the problems. It was never actually implemented. How do I know? I was asked if I could write such a script, which I said I would if the loader fix didn't work. I haven't written that script. Unless someone else has rebooted the server in the interim, it has not rebooted since back when we implemented the world leak fix.

2. This did not stem from "one person" reporting problems or having crashes. Nor were there any "vauge posts" involved. I was called in to do a server reboot for major lag. Proof? Server console was spewing "Cannot keep up" messages. After rebooting, that person then told me they could not connect to the server. I verified that by starting my FTB client and attempting to connect. Tell me again how this problem affects just one person?

2a. You're still bringing up CLIENT CRASHES. That's not remotely related to this problem. We're talking about SERVER crashes, and there's a major distinction between the two.

3. Well, off to the server log to see what went wrong. Lo and behold, I was greeted with a giant "EXCEPTION IN... blah blah blah". Okay, shit happens. Reboot server again. Server finishes loading mods, BOOM. "EXCEPTION IN....". Okay, time to read the crash report. Evidence in the log points at it being caused by loading a certain chunk. I disabled chunk loaders - and Voila! the crash is GONE and people can once again connect to the server and frolic about in the green fields of minecraft.

4. Everyone in this thread I have seen in this server more than once since the beginnign of April. Your argument there is invalid.

 

Color me skeptical if I don't believe you, but the server is most definitely not rebooting. Last logged reboot prior to yesterday I have is April 13, before that was April 2, and there are no plugins installed to that effect, nor is there anything in the startup chain/cron entries that kills/restarts the server.

If you don't believe me, excerpt from server log below. Timestamps are central, note the curious lack of reboots at 00:45 each night.