Tutorial Upbeat Giraffe - FAQ

Please make use of this thread rather than starting new threads for known issues. Muddying up this forum for common issues like port forwarding makes it difficult to determine what new issues are emerging.

Seems people are having some issues with the latest build so I figured I'd write this up after fiddling around with it last night.
Most of my answers will be coming from the perspective of running a public server without hardware limitations.

See my server post if you'd like to check out one of my three servers.

Changelog

Protocol 665 update notes
Protocol 666 update notes
Protocol 668 update notes

FAQ:

Q: Why aren't changes to my config file taking effect?

A: The Giraffe update has a separate "giraffe_storage" folder where the config file is being pulled from, not the one in the root of the Starbound folder. Make your edits to this one.

Q: How do I get admin to work?

A: See @metadept 's post regarding how to setup users in 668.

Q: Why can't I type anything into the server console?

A: This is by design. The server has to be controlled by admin commands that are accessible in-game. See the question above. To get access to more commands, you'll have to mod the game yourself or use a wrapper. Most of these have been abandoned since the last update. Supposedly StarryPy is being worked on by someone to get it working with Upbeat Giraffe, but I haven't personally looked into it.

Q: I'm getting a message on the server console: UniverseServer: Shutting down world CelestialWorld: blah blah blah due to inactivity. Why?

A: Do not worry yourself comrade.

This is a normal message and doesn't imply your server crashed or shut down. The server program has a single universe server and spawns multiple world servers for each planet players are active on. World servers get shutdown to save resources when nobody is in them. World servers get spun up automatically when you warp to them. This includes individual player ships and the Outpost.

Q: People can't connect to my server after the update, saying "Join failed!". What gives?

A: If you're using an application firewall (like Windows Firewall) and you switched to the 64 bit server or moved your installation, make sure you add the new executable to the list of allowed programs. If you're not behind a router that supports UPNP/has UPNP services enabled or you already have all of the proper ports forwarded and you are 100% sure you know how that works, try setting "upnpPortForwarding" in starbound.config to false.

Q: My server keeps crashing randomly. What's happening?

A: The latest build is on the stable branch but it's pretty far from actually being stable. If you plan to host for more than... I'd say about 8 players for an extended period of time, I'd expect crashes/lag every few hours. You'll just have to kick it when it dies or gets too laggy to be playable to get it back up and running. Setup scripts/services to do this for you if you can't monitor it constantly. There are a lot of guides on how to do this already.

Q: Why is everybody getting a "handshake failure" when they try to join? I can get in fine if I use my admin password but no one else can.

A: This appears to be some sort of bug with permanent bans. The only way I've found to fix this is to clear out the bans in starbound.config and restart the server. Leaving any part of them in the file doesn't appear to be effective anyway.

It would seem that at this time, the only safe way to ban someone is temporarily by adding the duration parameter at the end so it doesn't get written to the config file:

Code:

/ban "identifier" "reason" both "banDurationInSeconds"

The downside to this is that with how often the server crashes, bans will clear within a few hours at most no matter how long you set the timer.

Alternatively, you can add player IP addresses to your own firewall rules which works as a permanent ban. You can get a player's IP address by looking through the starbound_server.log files when they connect or disconnect.

Q: Linux or Windows? 32 bit or 64 bit?

A: From the testing I've done on both platforms, 64 bit handles more players smoothly (smoothly in the context of latency/visuals) but crashes pretty frequently. 32 bit chokes up quite a bit when you get past about 16 players regardless of your hardware and network connection. Even with very low CPU/RAM/Network/Disk utilization, my Windows 32 bit server eventually slowed to a crawl for people in-game after about a half hour. Linux is pretty clearly handling the game better on the server side at the moment, but If you already have a Windows box, I wouldn't go through the trouble of setting up a Linux box specifically for this game at this time. Server management is such a mess that it really isn't worth the trouble unless you plan to be on your server keeping an eye on people. The tile protection system is kind of a joke and without a better way to manage players, public servers are pretty much doomed to be trolled constantly.

The most stable for me so far is 32 bit Linux. I'm using a barebones install of Debian Wheezy, but I suspect any Linux distro will work fine. Most of the existing guides for Linux are for Ubuntu.

Q: Why don't server listing sites like Starbound Servers and Gametracker report the number of players on my server anymore?

A: The default configuration for Giraffe has the query server disabled, which is how those sites find out how many players are on your server among other things. Find:

Code:

"runQueryServer" : false,

and set it to true. I would also suggest setting "queryServerPort" to the same as your game server port just so you don't have to remember two different numbers. Make sure you update your profile on server listing sites to point to the new query server port. If you don't assign a separate query port, it's assumed to be the same as the game server port. This is okay to do because the query server is on a UDP port and the game server is on a TCP port.

Q: How does RCON work?

A: Based on information provided by AcidWeb, it appears that Starbound uses some implementation of Valve RCON. There are no simple out-of-the-box tools that support this to my knowledge. If you're familiar with one, feel free to share.

Q: What's causing the error: "Error: Fatal Exception Caught: (JsonParsingException) Error parsing json: expected JSON object or array at top level at 1:1"

A: This seems to have been caused by some change in how the file paths for dependencies are done in Giraffe, particularly the sbboot.config file. This can be pretty easily solved by specifying the working directory in scripts/batch files that launch starbound_server.exe to the correct folder (linux32, linux64, win32, win64).

Specifically for Powershell scripts using Start-Process, use the -WorkingDirectory parameter.

Q: How do I protect the spawn area using the new admin commands?

A: You would basically need to use a combination of the following commands:
/debug (shows the dungeonId where your cursor is)
/settileprotection dungeonId true|false (turn tile protection on or off for the specified dungeonId)
/setspawnpoint (set the spawn area for the current world to your character's current position)
This system is currently incredibly awkward to use. The dungeonIds aren't based on a grid system. It would seem that dungeonIds 65535, 65533, 65532 and 0 cover pretty much any given world, but in a rather haphazard manner.

If there's some more straight forward implementation of this you're aware of, feel free to share.

Q: People are getting around tile protection somehow. What's going on?

A: I can't monitor my server constantly but it appears some people with modified/hacked items can bypass tile protection, or they've found a way to bypass admin authentication. There's no way to prevent this at this time as the server won't check client assets against the server. The option to check asset digests in the config file doesn't appear to actually be implemented at this time.

If you have any other questions, I'll try my best to answer them in this thread.

Utility Scripts (currently only for Linux)

Deprecated since the ban bug is now fixed in 668, but keeping here for reference.

Below are some really crude bash scripts I put together for managing bans.
Might work on some others, but this is all I have for now.
I play my "I'm not a dev" card before you decide to mess with this stuff.
I'm not supporting these scripts. There's information online on how to implement this better than I did and it should be straight forward enough to follow the logic.

You should know what you're doing if you're going to use these. They're not user friendly and I don't know bash well enough to teach someone else how to use it.

Search logs for IP (findip)
Adjust the variable starboundHome to the correct location

Code:

#!/bin/bash

# Usage: findip <Name search>
# Doesn't handle colored names very well, search for a specific part of their name that shares one color

starboundHome="/home/starbound/server/giraffe_storage/"
tmp="/tmp/ipresults"

echo "Finding IP addresses that match the search:"

cat "$starboundHome"starbound_server.log* | grep -E -o "Client .+"$1".+<.+> \((25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[$
cat $tmp
echo ""
cat $tmp | grep -E -o "(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)" | sort -u
rm $tmp

Example output:

Code:

findip Al
Finding IP addresses that match the search:
Client 'Al' <14> (82.29.103.xxx
Client 'Al' <66> (82.29.103.xxx
Client 'Alexis' <38> (109.202.86.xxx
Client 'I-Alen' <56> (37.230.228.xxx
Client 'The Alpaca god' <127> (50.155.122.xxx

109.202.86.xxx
37.230.228.xxx
50.155.122.xxx
82.29.103.xxx

Ban manager through iptables (ban)
since Starbound's bans are borked

Code:

#!/bin/bash


# Usage:        ban                    |      Shows current bans and refreshes the chain
#              ban <IP address>        |      Adds given IP address to the chain

banFile="/home/starbound/bans.txt"
tmp="/tmp/ban"

if [ -n $1 ]; then
        echo $1 >> $banFile
        echo "Banning: $1"
fi

sed 's/^ *//; s/ *$//; /^$/d; /^\s*$/d' $banFile > $tmp
sort -u $tmp -o $banFile
rm $tmp

iptables -F STARBOUND
iptables -X STARBOUND
iptables -N STARBOUND
iptables -D INPUT -j STARBOUND
iptables -A INPUT -j STARBOUND

while read ip; do
        iptables -A STARBOUND -s $ip -j DROP
done < $banFile

iptables -A STARBOUND -j RETURN

iptables -L -n

SUPER BOAT (incomplete)

The basic idea behind building your own cool protected starting are and having the home planet reset everytime your server restarts.

Building your own super boat (or whatever cool starting area structure) that is recreated every time the server starts is fairly straight forward.

Find out what world your server uses as the starting planet for all players. They're named after the coordinates you see them under in the galaxy map (you can now use /whereami in 668 to get this as well) with some other numbers tagged on to the end to help you narrow it down. If you just log on to your server then warp down to that world and make some quick changes like adding a few blocks, then you can use the last modified timestamp on the files to find which one is the one you were just on.

If your server is running, stop it gracefully (Press CTRL+C on the server console in Windows or send SIGINT to the process in Linux) and delete the world file for that planet, including any that end with the extension 'fail'.

At this point I would put a password on your server so you and your fellow builders can work without getting interrupted.

The next time you start your server, it will regenerate that world. Now go find a good place to start building your structure. I chose to put mine in the sky as it's easier to handle working with a blank slate when doing tile protection.

If you do choose to go with the floating structure, I would put it above a randomly generated dungeon like a mining shaft so you can protect the area right below your structure as well.

Once you're finished building, you can put some starting items (like core fragment ore) or instruments into storage areas inside.

Now enable tile protection for that building(/settileprotection 65532 true) and start testing to make sure tiles and placed objects can't get destroyed. It's very unlikely you'll get this right on the first try and you'll have to do some planning based on how tile protection currently works.

In order to prevent people from putting new tiles down on top of your existing tiles, you have to add a back layer as far as you want to prevent placement of objects in the foreground. Large objects will be able to get knocked down if the entire space they occupy doesn't have a backing wall.

The only alternative I found to this is to protect dungeonId 65535 as well, but this makes building anywhere on the starting planet pretty much impossible and as well making digging more difficult as well. If you want a completely static starting planet though, this is probably the way you'd want to go.

Once you're all done, make sure you set the spawn point on the starting planet to the desired spot with /setspawnpoint (which will the warp in area to your current location).

When you're done, shut down the server gracefully (Press CTRL+C on the server console in Windows or send SIGINT to the process in Linux) and wait for it to completely end the process. Copy the world file for the starting planet and just keep a copy of it outside of the universe folder. Keep a note of the file name as this won't change unless you completely wipe out your universe folder (to my knowledge).

In whatever launch/startup script you have for the server, before it starts the server process, have it delete anything with the file name of the existing world and copy the world file you saved earlier back into the universe folder.

You should now have a cool starting area that is completely protected and a home planet that people can mine and play around in anytime the server restarts.

 

They used generic implementation of RCON.

It have to be enabled in config like QueryServer + separate password and port.
After connecting to it ingame commands are available - but without / at the start.

 

Thanks man, I was getting the bans/handshake error and I had a feeling it was something to do with this but couldn't find any information or anyone else experiencing this problem so I was unsure. Its annoying that I cant truly ban trouble makers at the moment in this case. To be honest the only trouble makers I'm getting are people that have spawned in stupid weapons that fire projectiles that do explosion damage to other players regardless of if the planet is pvp-enabled or not or enough bombs to destroy complete planets. This wouldn't be a problem if server-side inventories were in...

I would love to see the Starbound developers start taking the server software/features seriously as little to no testing seems to have been done on this side of things. How do you expect us to host and promote your game when the server software is broken and people are running a mock in our servers? Most of these problems seem to be down to developers seemingly ignoring server hosts and not providing the necessary tools and features to moderate a server properly.

I want to host this game, enjoy the game with others, promote this game and sing its praises but at the moment I'm finding it difficult and frustrating when my questions go unanswered and when our server software is so direly unstable.

Again, massive thanks Nordan, at least someone shaires my pain, and is striving to help others. Wish we could get this seen by the Devs though, I doubt they even read the multiplayer board...

 

Not sure what a generic implementation of RCON is. It's not a protocol or a framework with a base client, it just stands for remote console. A lot of different software implements the concept of RCON differently. The settings are there in the config but the devs haven't put out anything on how to interface with it as far as I could see. If you have any information on it, feel free to share. I have it enabled but haven't found any obvious way to connect to it.

My understanding is that they now have some deal with Multiplay for servers and I suspect they're making a move towards what DayZ Standalone is doing, essentially butting out independent hosting or giving preferred hosting services the leg-up in terms of game server management in a mutually beneficial deal.

Where Starbound server hosting is right now, my opinion is that they shouldn't have it if their only response of the criticism is going to be that it's in beta. What admins have to control their servers at this moment is on par with tools created from reverse engineering and hobby projects. See how many wrappers and tools have been abandoned since the last update. Fully agree with your sentiments on the state of the dedicated server. I like the concept of the game but I'm not going to look past its faults when they've already got the money from sales.

 

My personal experience with multiplay was horrendous, they are a god awful hosting provider and their services and support were just the worst. If the game did go down this kind of route I think it'd be a tragic loss of a game that had so much potential. I completely agree that the "oh it's in beta" excuse is becoming just that for every little problem that the game is facing. Fact of the matter is this should have been tested before release. It is in no way, shape or form a "stable" release. Just hugely disappointed really.

 

I've tried hosting ~6 players on a Linux VPS with 3GB of RAM (Debian 7 x64, linux64 server, it would use 2.5GB of RAM right after launching) and this would crash every hour or so with the message "Killed" in console and this in dmesg:

Code:

[18955147.815276] Out of memory in UB 4138: OOM killed process 18678 (starbound_serve) score 0 vm:4418032kB, rss:3102876kB, swap:5472kB

I've switched the world over to my Windows 8.1 64 bit PC with 8GB RAM and so far it hovers around ~500MB of RAM usage, any reason why the linux server uses such a ridiculous amount of memory?

 

@Goz3rr - Sadly this is normal for 64bit binary. It is just unstable and have memory leaks.

@Nordan - Same/similar as VALVE RCON. For example you can use python-valve library to make connection.

 

Have you tried running the 32 bit Linux server? Couldn't tell you why it's hogging your RAM, might be a memory leak in the 64 bit binary. My server has 32 GB of RAM but the 64 bit Linux server would crash repeatedly within 10 minutes, although it was handling over 20 players.

I haven't seen the Windows binaries use more than 2.5 GB of RAM even when there are about 30 players on, but you see performance issues regardless even if your CPU utilization isn't that high. I suspect the server doesn't handle multithreading very well once you hit a certain number of players.

 

I switched to linux32 and it hasn't ran out of memory yet, however i've had this happen twice:

Code:

Info: UniverseServer: Shutting down world CelestialWorld:-845942946:-93583624:21961937:6:1 due to inactivity
Info: UniverseServer: Loading world db for world -845942916:-93583624:-122358643:6:1
Info: UniverseServer: Loading world db for world -845942916:-93583624:-122358643:6:1
Error: WorldServerThread exception caught: (DBException) Error, incorrect leaf block signature.
./starbound_server(_ZN4Star13StarExceptionC2ESs+0x6f) [0x8a27d4f]
./starbound_server(_ZN4Star13StarExceptionC2EPKc+0x22) [0x8a28102]
./starbound_server(_ZN4Star13BTreeDatabaseINS_9ByteArrayES1_E10LeafReaderC1EPS2_j+0xe1) [0x89dc911]
./starbound_server(_ZN4Star13BTreeDatabaseINS_9ByteArrayES1_E9BTreeImpl8loadLeafEj+0x37) [0x89e0ca7]
./starbound_server(_ZN4Star10BTreeMixinINS_13BTreeDatabaseINS_9ByteArrayES2_E9BTreeImplEE4findERKSt10shared_ptrINS_16SimpleBTreeIndexIS2_jEEERKS2_+0x3e) [0x89e1eee]
./starbound_server(_ZN4Star14SimpleDatabase4findERKNS_9ByteArrayE+0xb3) [0x89d9d53]
./starbound_server(_ZN4Star12WorldStorage10loadSectorERKNS_6VectorIjLj2EEENS_15SectorLoadLevelE+0x2db) [0x8406b8b]
./starbound_server(_ZN4Star12WorldStorage10loadSectorERKNS_6VectorIjLj2EEENS_15SectorLoadLevelE+0x1b1) [0x8406a61]
./starbound_server(_ZN4Star12WorldStorage14generateSectorERKNS_6VectorIjLj2EEENS_21SectorGenerationLevelE+0x3f) [0x840708f]
./starbound_server(_ZN4Star12WorldStorage6updateEv+0x93) [0x8407273]
./starbound_server(_ZN4Star11WorldServer6updateEv+0x14d4) [0x83dd444]
./starbound_server(_ZN4Star17WorldServerThread6updateEv+0x1bb) [0x840215b]
./starbound_server(_ZN4Star17WorldServerThread3runEv+0x26b) [0x84026eb]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x1a) [0x8a3011a]
/lib/i386-linux-gnu/i686/cmov/libpthread.so.0(+0x6efb) [0xb7711efb]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(clone+0x5e) [0xb75fe10e]
Info: UniverseServer: Shutting down world ClientShipWorld:f0f540a5fc43d04a79f78cc7d9909ef4 due to inactivity
Info: UniverseServer: Client 'Skyrox' <4> (86.158.21.141:60230) disconnected
Info: closing 86.158.21.141:60230 (13)
Info: NetSocket read loop shut down: (SocketClosedException) Connection reset by peer 86.158.21.141:60230 (0)
Info: accept from 86.158.21.141:61026 (14)
Info: UniverseServer: Connection received from: 86.158.21.141:61026
Info: UniverseServer: client connection made from 86.158.21.141:61026
Info: UniverseServer: Assets digest check is disabled during client connect
Info: UniverseServer: Sending Handshake Challenge
Warn: UniverseServer: Logged in account  as player Skyrox
Info: UniverseServer: Client 'Skyrox' <7> (86.158.21.141:61026) connected
Info: UniverseServer: Loading ship world received from client <User: Skyrox>
Info: UniverseServer: Loading world db for world -845943031:-93583585:16928618:8
Info: UniverseServer: Shutting down world CelestialWorld:-845943031:-93583585:16928618:8 due to inactivity
Info:  <Lucy> is there a list of all crafting recipes I know?
Info: UniverseServer: Loading ship world received from client <User: Cpt Flamehead>
Info:  <Goz3rr> dont think so
Error: UniverseServer: error during worker pooled job: (DBException) Error, incorrect leaf block signature.
./starbound_server(_ZN4Star13StarExceptionC2ESs+0x6f) [0x8a27d4f]
./starbound_server(_ZN4Star13StarExceptionC2EPKc+0x22) [0x8a28102]
./starbound_server(_ZN4Star13BTreeDatabaseINS_9ByteArrayES1_E10LeafReaderC1EPS2_j+0xe1) [0x89dc911]
./starbound_server(_ZN4Star13BTreeDatabaseINS_9ByteArrayES1_E9BTreeImpl8loadLeafEj+0x37) [0x89e0ca7]
./starbound_server(_ZN4Star10BTreeMixinINS_13BTreeDatabaseINS_9ByteArrayES2_E9BTreeImplEE6modifyERSt10shared_ptrINS_16SimpleBTreeIndexIS2_jEEERNS5_10ModifyInfoE+0x76) [0x89e99c6]
./starbound_server(_ZN4Star10BTreeMixinINS_13BTreeDatabaseINS_9ByteArrayES2_E9BTreeImplEE6modifyENS5_11DataElementENS5_12ModifyActionE+0x250) [0x89eacf0]
./starbound_server(_ZN4Star13BTreeDatabaseINS_9ByteArrayES1_E6insertES1_S1_+0x6f) [0x89eb2cf]
./starbound_server(_ZN4Star14SimpleDatabase6insertENS_9ByteArrayES1_+0x3b) [0x89dabfb]
./starbound_server(_ZN4Star12WorldStorage16setWorldMetadataERKNS_16VersionedVariantE+0xa2) [0x8404f92]
./starbound_server(_ZN4Star11WorldServer13writeMetadataEv+0x7af) [0x83d62ff]
./starbound_server(_ZN4Star11WorldServer4syncEv+0x10) [0x83d6710]
./starbound_server(_ZN4Star17WorldServerThread4syncEv+0x27) [0x8400aa7]
./starbound_server() [0x838ad39]
./starbound_server() [0x8a21e6f]
./starbound_server(_ZN4Star10WorkerPool12WorkerThread3runEv+0xda) [0x8a225ba]
./starbound_server(_ZN4Star10ThreadImpl9runThreadEPv+0x1a) [0x8a3011a]
/lib/i386-linux-gnu/i686/cmov/libpthread.so.0(+0x6efb) [0xb7711efb]
/lib/i386-linux-gnu/i686/cmov/libc.so.6(clone+0x5e) [0xb75fe10e]
Info: UniverseServer: Shutting down world CelestialWorld:-845942916:-93583624:-122358643:6:1 due to inactivity
terminate called after throwing an instance of 'Star::DBException'
  what():  Error, incorrect leaf block signature.
Aborted

A player loads a (new?) world, nothing shows up and when they try to warp back to ship the server crashes instantly. Warping to the same planet after the restart works fine

 

I've had that happen myself quite a few times. It's the only error I've seen crashing the 32 bit Linux server so far. I think it has to do with people warping around too quickly, but I haven't tested it extensively. I don't think there's any setting or variable for us to adjust to prevent this, most likely something that just needs to get patched.

 

Bumping this as people keep creating new posts about issues mentioned here.

 

Bumping again. If the devs don't fix this, then a serious modder like me is going to have to do something about it.

 

Hello OP!

A wonderful and thought out list of Q's and A"s


I have

Code:

"queryServerPort" : 21025, 
"runQueryServer" : true,

And yet gametracker cannot scan the server and server status has NA for p layers.

Any thing else i can try to fix it?

 

Hi pandee,

Keep in mind the query port is a UDP port. Your firewall/port forwarding rules will need to include 21025 for UDP as well as TCP if that is your game port as well.
Also if I recall, Gametracker assumes a different port other than the game port as the default for the query server. Verify your query port on Gametracker is set to 21025.

 

Yes i have it set to 21025 =]

Anything else?

 

You didn't mention if you checked whether your firewall/port forwarding rules were configured for UDP traffic on port 21025.

 

Hey guys, Ive found an easier way to deal with ip bans, (may only work with windows servers)

Bypass the server ban list altogether and add any unwanted ips to your windows firewall blocklist.

 

Added some bash scripts for managing bans through iptables.

Wouldn't suggest using them unless you know what you're doing. bash is for Linux.
Shouldn't be that hard to convert the logic to something that works in Powershell for Windows folks.

 

HI.

I'm searching for a system for Win 2012 that restart server after he crash.
How can it be done ? Several methods ?

Cheers and tahnks for the help.

 

Simply open the advanced firewall page and in outbound/inbound connections make a new rule. Make it for a port, make that port 21025 and be sure to check UDP not TCP, save and you're all done.