Nocom – 2b2t Minecraft server exploit using Monte-Carlo localization

https://github.com/nerdsinspace/nocom-explanation/blob/main/README.md

by cyber_kinetist • 5 years ago

208 80 5 years ago

github.com

cyber_kinetist• 5 years ago

Here's the FitMC Youtube video that first announced the exploit: https://www.youtube.com/watch?v=elqAh3GWRpA

A 2b2t wiki page about the exploit: https://2b2t.miraheze.org/wiki/Nocom

The teaser video: https://www.youtube.com/watch?v=3ayxeruAan8

323• 5 years ago

I second the first video, it's extremely well done, and as someone who never played minecraft I almost understood what happened.

bastardoperator• 5 years ago

"2b2t ,the oldest anarchy server in Minecraft" - FitMC (every video)

H8crilA2• 5 years ago

This is from the same guy who coded Baritone:

https://youtu.be/CZkLXWo4Fg4?t=78

Extremely useful on anarchy servers, in fact the entire current "era" of anarchy is called "The Automation Period" mostly due to Baritone.

leijurv26• 5 years ago

I actually helped convince Sato in Feb 2020 to call it "automation period"; some other ideas were "age of fragmentation" (referring to how group alleigances were becoming split, with players in more than one group at a time) or "age of the algorithm" (referring to the YouTube algorithm promoting the server).

leijurv26• 5 years ago

Surprised to see this posted again, I wrote this, feel free to ask me anything!

ucarion• 5 years ago

I'm curious what kind of tooling you developed as part of this process. It seems, from the linked YouTube video, like individuals were actively monitoring the data produced from the exfiltration attack. How did you go from the raw data to producing something that people could usefully monitor?

leijurv26• 5 years ago

Sure!

For monitoring on a meta-level, we used Prometheus and Grafana. Here's what that looked like near the very end: https://www.dropbox.com/s/ktlvvtmppq64u8o/trimmed%20grafana%... Since the ratelimit was heavily in place by then, the numbers are pretty low. For example, the headliner "blocks per second" was well over three thousand a few weeks before then.

One of the most common queries for actual usage though, was "what bases in the last 24 hours (or whatever) have the most chests currently". We'd then travel to these locations to borrow items. Here's that query https://pastebin.com/fjhHGSYz Roughly, it grabs chunks with chests in that timeframe, then uses a recursive Postgres CTE to traverse the https://en.wikipedia.org/wiki/Disjoint-set_data_structure that we used to implement DBSCAN, from the leaf to the root. Then it groups by the root, which in practice means grouping chunks together by which ones semantically have been determined to be part of the same base, then makes a report by quantity per base.

There was also a web UI that showed all active tracks. You could click any player and see their travel history, the route they had taken to get to that location. As well as clicking any cluster and seeing the usernames associated with it. We used this a lot to grab what members were present at a given item stash or base, to make sure we weren't stepping on anyone's toes.

In terms of what it took to make this happen, the stars of the show were probably the main Postgres, which was hand-tuned on top of a hand-tuned ZFS, on top of a NVME drive that was IOMMU passthrough'd to this VM for maximum performance. As well as https://github.com/OpenHFT/Chronicle-Map, which really helped out. We needed a mapping from block coordinate to a bit of data about that location, such as the last timestamp we checked it, what Minecraft block was there last, whether the block is interesting, uninteresting, different from seed, etc. High frequency trading software was perfect for this use case, because it greatly reduced GC pressure on the JVM, since the data was stored off-heap. It ended up at around (roughly) fifty million entries, served hundreds of thousands of requests for block data per second, and requested historical data (from the Postgres) at about thirty thousands SELECTs per second (from a table with over ten billion rows). The reason why this needed to be so fast, is that the use case is downloading bases. Sometimes someone would log in to 2b2t at their base, after having not played for weeks at a time (so the data is in Postgres, not the map in RAM), we wanted to be able to pick up right where we left off on downloading what they had built. This means it needs to get back up to speed with their entire base, as fast as possible, in the seconds before they move away to other chunks, or log off the server. We are downloading bases one click at a time, and bases are huge. The area that is loaded in by a player is 144 by 144 by 256. That’s 5 million blocks. And there are about 250 players online on 2b2t at any given moment in time. That’s well over a billion blocks that we could click on, and learn which Minecraft block they are. But, which locations should we use up our clicks per second budget on? It focused on areas that are being changed rapidly, and areas that are different from the base vanilla terrain. Those are certainly bases, large Minecraft builds. We do an expanding paintbucket pattern that expands in 3 dimensions to whatever has been built there. However, we have to do this quickly, and we have to pick up where we left off. Someone might be flying around their base with an elytra, at very fast speed. We would love to scoop up the terrain that they’re flying over. If the base is particularly interesting, someone flying over a certain area could have had about twenty Minecraft accounts spread all over 2b2t’s map suddenly refocus all of their left clicks on that one area, to grab all the interesting neighbors that it wasn’t able to last time. We peaked at about three thousand block clicks per second, averaged over weeks. So this was a massive data loading and management problem.

fffobar• 5 years ago

Tracking was completely broken by any teleporation, right? Such as enderpearl loading or a respawn at the bed.

Also, how many login sessions would it take to associate a track with an account?

leijurv26• 5 years ago

Correct, loading a pearl and respawning would throw us off the trail. This is why our aim was to follow the initial travel to the base location, rather than a future pearl TP. The easiest advice in order to throw off nocom is to travel in a straight line in the overworld for perhaps 10k blocks, set a bed quickly, then continue another 10k blocks, then /kill to the bed, then go perpendicularly for a render distance or two before the chunks unload at your death location. Even with some hypothetical tracking improvements (that never actually were written), such as walking back the last 10 minutes whenever a track goes cold, that would still throw it off.

Generally 2 or 3. 1 point of association was given whenever a track went cold, split up among however many players logged off within -6 seconds to +1 seconds of that timestamp (because of chunks unloading on a 5 second schedule). 1 point could be a false positive, but 2 or 3 would be reliably trustworthy.

ThouYS• 5 years ago

what's your take on university as opposed to learning by doing? how do they complement each other?

leijurv26• 5 years ago

I am not done with university yet, so I don't speak confidently. But I'm afraid I have no special perspective, I just echo what I hear a lot: Learning by doing is probably the most important, but structured courses provide a theoretical and practical foundation that fills in a lot of important gaps that I would have had if I were only self taught.

ajkjk• 5 years ago

Why did you do all this? I mean.. it's a ton of engineering, for what? For fun?

leijurv26• 5 years ago

Yes absolutely, for fun! Of course, it's also fun to have a secret project that no one knows about, to get an edge. 2b2t is practically the only Minecraft server that allows any kind of hacking your client (without getting banned), while also having an active community of meaningful size, so that was the clear target.

rasz2• 5 years ago

griefing

Btw Mirai botnet also came out of minecraft.

barneygale2• 5 years ago

Hello! I'm the original owner of the 'BibleBot' account/bot. Glad to see you put it to good use :-)

leijurv26• 5 years ago

Lol! Sorry for getting it locked, it is reversible if you still have the email. Mojang, it appears, didn't like spam login from a different cloudflare warp IP every few minutes...

hansihe• 5 years ago

Fun to see you around here as well!

phgn2• 5 years ago

This whole project is incredibly impressive, good job!

How much time would you say went into this both from yourself and the others? Probably quite a lot over the years? (because again, it's very impressive)

leijurv26• 5 years ago

Hard to estimate, probably a few hundred hours from me, and a comparable amount from others. As a random data point, I sent 53000 messages in the group chat developing it. ¯\_(ツ)_/¯

tommsy64• 5 years ago

I'm curious how you developed the "very cute headless" Minecraft client. Did you use the Forge build tools? How did you go about ripping out the rendering/keyboard stuff?

leijurv26• 5 years ago

Yes, Forgegradle was used. Frankly, not much interesting here, just a whole lot of work going through all the Minecraft code, and using Java reflection where possible and JVM bytecode injection otherwise. There's a lot of access transforming, using theUnsafe to allocate classes without calling their constructors, etc. As a random example that I remember, one of the last things to patch was that we had it working, but when you got kicked from the server, it would crash. This is because it would try to create a currentScreen GUI for the "disconnected from server" message, which for some reason was long enough that it was considering whether to wrap the text to the next line, which needed something like a font size, or maybe the window width, which crashed since there is no window or font.

potatoman22• 5 years ago

Did you use some form of Bayesian statistics to choose which chunk to query?

leijurv26• 5 years ago

Nothing so advanced, I'm afraid. The Monte Carlo system just creates a thousand points (2d position and velocity), and simulates them at 1 generation per second, culling points that are unlikely and reproducing points that are likely based on observations (clicking chunks and hearing back "loaded" or "unloaded").

snowycat• 5 years ago

Really impressive project! How did you learn all the things that you needed to make this? Are there any particular books or resources you would recommend? Also, how much did this entire project cost? I'd imagine processing that much data and storing it wasn't cheap.

leijurv26• 5 years ago

Mostly learned as I went! I had never before implemeneted Monte Carlo localization, or DBSCAN, or a disjoint union find; it was a lot of research and trial and error. I'd honestly recommend Wikipedia! I also read a lot of blog posts such as on data clustering in order to decide how to write an algorithm to tell what is and isn't a single contiguous base.

The main expense was priority queue, which is a $20/account/month payment to 2b2t to be able to join the server faster. Probably nearing a thousand dollars on that. We lightened this by creating a proxy system where other builders could join through our headless, and it would add nocom exploit packets from their account to 2b2t, and peel off the responses on the way back. They got a more reliable experience with less disconnects, and we got some extra capacity for free (completely unbeknownst to them of course haha). By the end, all the overworld accounts were being paid for and used by other people, I only had to cover the 2 accounts in the nether.

About $300 on a new 2TB nvme drive for postgres. An unknown amount on electricity. Other than that, all the big processing ran on an existing homelab server (thanks to fr1kin). About $150 on digitalocean for a droplet to run the headless minecraft with minimal ping to 2b2t ($10/month). Backups were just WAL rsyncs to postgres replicas ran (again on homelabs) by myself and a few others.

gwern2• 5 years ago

Why did it take them so long to discover the issue when you and another group were practically DoSing them every second for years on end (and they kept lowering the packet limit)? Didn't they ever wonder what it was all for or what the outgoing packets contained?

leijurv26• 5 years ago

This is one of life's great mysteries. Given the hosting provider for 2b2t, we know pretty much for a fact that the packet spam from 2018 and 2019 (before I knew about the project ;)) would have cost tens of thousands of dollars in bandwidth overages per month. Yet, he never added a packet limiter until an unrelated exploit could cause other clients to crash due to spamming an offhand swap sound effect. It took the exploit going fully public, plus another 36 hours or so, for it to actually get patched. ¯\_(ツ)_/¯

heartbeats• 5 years ago

How accurate a map did you end up with?

leijurv26• 5 years ago

For tracking, the spatial resolution was 1 chunk, which is 16 by 16 blocks, and the temporal resolution was 1 second. So, we have a few billion rows of tracking data for every player, localizing them to roughly that distance, roughly once a second.

For downloading bases, it is an essentially perfect recreation. Some things are missing, such as the color of banners, but for 99% of blocks, just setting the correct blockstates will recreate the build. Some disconnected sections of the build might not be found by the paintbucket floodfill algorithm though, so floating parts could be missing. We counteracted this by having a few random blocks in the chunk be checked on a schedule, so eventually it would get everything.

the_gipsy• 5 years ago

This is brilliant!

Ah, gives me nostalgia of civcraft conspiracies...

thegjp210• 5 years ago

good to see some other civcraft alumni on HN. What an experience...

iFred• 5 years ago

Probably the only server experience that really felt like you had an entire world and society to explore and consequences to deal with.

I spent way too much time trying to figure out a way to scrounge up Ghasts and TNT to level Orion.

meibo• 5 years ago

What confuses me about this is that no one else found it?

Being somewhat experienced at game networking and writing server code with this type of exploit(e.g. sending a packet to do something at the other end of the map and processing the result) in mind, this would be something I either would have checked for in the first place if performance would have allowed for it, or would have kept in the back of my mind in case I hear about a possible exploit that takes advantage of it.

leijurv26• 5 years ago

One reason might be that this only existed in Paper, which is a downstream fork of Spigot, which is (sort of) a downstream fork of CraftBukkit, which is a set of patches to the official minecraft server.

Another reason might be that this only does anything if there are many players on the server, and you hit a region loaded by someone else. Might be difficult to think of that. You'd never come across it if testing in singleplayer.

But in essence I agree, it's surprising that no one else found and reported this years ago!

Sesse__• 5 years ago

What confuses _me_ about this is that nothing seems to say what it _is_. It's like there's some giant practical joke I'm not a part of, and the README.md just throws off heatmaps and wants me to watch a 24-minute YouTube video. Is everybody supposed to be a Minecraft expert? Is a simple paragraph of what this is all about (at the very beginning of the file) too much? :-)

leijurv26• 5 years ago

There's a section called "The entire thing summed up in three sentences" in the linked GitHub README.

bobsmooth2• 5 years ago

"You could punch any block coordinate in the entire map, and 2b2t would either:

-Tell you what block is placed there, if the chunk is currently loaded

-Or ignore you, if the chunk is not currently loaded.

2b2t would immediately reveal whether or not any chunk coordinate that you ask about is currently loaded. If it is currently loaded, it tells you which specific block is there at the coordinate you asked about."

This allowed the hackers to find every player's coordinates and where their bases were.

rasz2• 5 years ago

Its a typical case of Software developer not validating player input 'because thats too taxing on the server'. Minecraft server was programmed to reply to all, even stupid or impossible queries.

itronitron• 5 years ago

For most multiplayer MC servers, this sort of exploit isn't worth the time because there are commands available to players to find nearby players and their bases (so they can check out each other's builds or participate in the local economy.)

cityofdelusion• 5 years ago

If anyone else was lost like myself, here is the rundown. The linked github readme assumes prior knowledge of 2b2t.

This is an exploit for a very old Minecraft server (called 2b2t) where griefing (destroying other creations) and hacking is basically allowed. This kind of environment means that groups of players are incentivized to hide their projects out in the middle of nowhere, hundreds of thousands of blocks away from the world origin coordinates. The exploit allows a client to ask the server if data chunks at arbitrary coordinates have player activity in them, opening the server up to a brute force search to locate well-hidden bases.

Basically, a group got spy satellites in an anarchy Minecraft server where remaining hidden is critical to not being destroyed.

kingcharles• 5 years ago

Thank you. I had no idea what was going on. This is actually pretty neat once you explain it.

Bluecobra• 5 years ago

This is a pretty clever exploit. I played on this server off and on between 2013-2016. The server is/was an acquired taste. The biggest challenge was to escape the the hellscape around spawn, once you get past that it was pretty quiet, I don’t ever recall running into anyone else unless I went back to the nether hub near spawn. It was really fun exploring old builds, the griefed ones were like old ruins. I liked to read the signs that were left behind and leave my own. I used to have use sone map mod and would note all the neat things I came across.

I admit that I also used an x-ray mod to detect things like chests, because why not? I wonder if any of my bases were ever discovered. Most of my ill gotten gains were hidden away in a bottom of a mineshaft in the middle of nowhere to help avoid detection.

Once the queue started started to get long I stopped playing.

rvnx2• 5 years ago

In Minecraft you can dig blocks.

When you click to dig a block, it just calls the RPC method CPacketPlayerDigging with the coordinates of the block.

The server replies "Block of type ABC is too far".

So, of course, you now know that at (X,Y,Z) you have a block of type "ABC".

This RPC has no rate-limiting, so yes, you can know the whole map.

The fact that you can dig any block on the map, seems like a shortcut taken by developers of the server to save time or resources.

I'd even say it's almost intended behaviour.

Tech details: https://2b2t.miraheze.org/wiki/Nocom

leijurv26• 5 years ago

Not quite a shortcut, it was actually to patch a bug in vanilla minecraft that causes desync, see https://news.ycombinator.com/item?id=29621262 And it may seem silly that it tells you that it's block type ABC, but the reason is to tell your client to put it back as block ABC, instead of air. (in other words, it has to "undo" your client claiming to have mined that block [which replaces whatever it is with air]).

TheJoeMan• 5 years ago

I learned some of the Monte Carlo techniques in my medical imaging classes, it’s very interesting the data one can fish out with a good-enough graph.

I also find the adaptive location “pinging” quite interesting, reminds me of the tech behind the F-22 active radar.

bobsmooth2• 5 years ago

My favourite part of this has to be the social engineering used to get the exploit patched in.

kregasaurusrex2• 5 years ago

The code used in the exploit itself itself looks rather harmless, as it was created to be a mitigation against finding players based on previously loaded chunks. https://github.com/PaperMC/Paper/blob/ver/1.12.2/Spigot-Serv...

leijurv26• 5 years ago

That is not correct, this patch is meant to prevent block break desync, which is when your client and the server disagree about what blocks are placed in the world. If the server discards/ignores anything your client sends, the server MUST reply telling the client to undo that action, otherwise desync occurs. (e.g. "i have broken this block" -> the client is too far to reach that block -> the server has to reply with "set that block back to air").

The original version of this patch https://github.com/PaperMC/Paper/blob/d934dcc1e3e8b1ac34acef... fixed that desync, but it allowed you to force the server to generate an unlimited amount of chunks, which is an expensive operation.

You could only find players using this patch after it was modified to only reply if the chunk was already loaded. But at no point in time was this patch able to reveal if a chunk had been previously generated or not, that is a separate unrelated issue in which Spigot will reliably send a partial chunk on initial generation, but never for a previously generated chunk.

Take a look at https://2b2t.miraheze.org/wiki/Nocom

kregasaurusrex2• 5 years ago

Thanks for the clarification!

w-ll2• 5 years ago

Why would the `mineblock` packet not be clamped to at least the render distance or more realistic the chunk the sender is in and the 8 adjacent chunks. I know i'm learning about this in hindsight and it's obvious.

leijurv26• 5 years ago

Yes, so Minecraft is exactly that. Any "I have mined this block" further than 6 blocks away from the player is ignored. The problem is that on a laggy server, an honest client can end up sending this, if the server lags out for more than 6 blocks of walking your player forward. So, Paper added a patch to undo these disallowed mines, setting them back to what they were. But this happened at any radius, and could be used to lag the server by making it generate any region. Then they patched it again and made it only reply if it were in a loaded chunk. The mistake was that it replies if the coordinate is in a chunk loaded by any player, not just by your player. :) See https://news.ycombinator.com/item?id=29620852

w-ll2• 5 years ago

Hey thanks for your reply. I get how the Paper patch allowed this new exploit, but Im still curious to why Paper never even had the range check. Even for lag compensation it shouldnt be more than a chunk or 2 range.

Im actually reminded to the latest log4j exploit (main reason i aint actually gonna run Minecraft for a while) and even the Heartbleed exploit a few years back.

Did yall assume the patch to Paper would just return if the chunk was player loaded, and they wouldn't add any other sane bounds checks?

If the story from the FitMC guy is true, yall guessed/hoped that patch would be what it was. Giving yall a new leak of information without actually fixing the problem because if it was loaded you could still get all the block info in that chunk.

Pretty awesome from where i'm sitting. \(ᵔᵕᵔ)/

howenterprisey• 5 years ago

https://2b2t.miraheze.org/w/index.php?title=Space_Valkyria_3... Wow... I love this edit so much.

leijurv26• 5 years ago

Thanks!

Nocom – 2b2t Minecraft server exploit using Monte-Carlo localization

Related Stories