Closed Persistent Crashing on Specific Worlds

Lately, when playing on Sme4gle's Starbound server, I have had the game conssitently, and silently, crash when on specific worlds No exception error, no error message at all. It just closes to the desktop.

I have noticed that when these crashes occur seems to vary - sometimes I can run around a couple screens on the surface, other times the crash occurs the moment I beam down. This leads me to believe it's monster-related.

Looking at Starbound.log, there's a world-stop packet right before the crash occurs, and then there's a bunch of error strings. Oddly enough, it didn't catch those the previous the couple times this happened.

I've attached Starbound.log, just in case there's something in it that I missed. Had to change the extension, since the forums apparently don't like files with the extension *.log on 'em.

 

Does it happen in single player on the same planets?

 

I would test that, but unfortunately that system doesn't exist in my single-player universe, due to OS differences between the server and I.

EDIT: I'm getting Sme4gle to post the server logs shortly. Should shed some light on the issue.

 

Server in question is my server, which is running Ubuntu Server 14.04.2 LTS (x64). I have managed to create a log of the server for Nerva which states that the world db for the world Nerva is trying to access could not be loaded.
This occurs on a pretty large amount of planets, And has been happening before on my windows server aswell (Windows 7 Professional x64).

[Edit:] On my Windows server this happened on a different amount of clean generated universes, did not try this on the linux server though.

 

Thank you for providing the server log.

Code:

[21:44:33.525] Warn: SimpleDatabase at device: ./../giraffe_storage/universe/universe.chunks was not closed cleanly
[21:44:33.525] Warn: Recovering SimpleDatabase at device: ./../giraffe_storage/universe/universe.chunks
(..)
[21:44:33.668] Warn: SimpleDatabase at device: ./../giraffe_storage/universe/playerCodex.db was not closed cleanly
[21:44:33.668] Warn: Recovering SimpleDatabase at device: ./../giraffe_storage/universe/playerCodex.db

Did the server crash or was it killed to exit? http://community.playstarbound.com/index.php?threads/best-practice-to-shut-down-a-server.34522/

Code:

[21:56:02.775] Error: UniverseServer: Could not load world db for world -111727825:81710841:-272813499:2:5, removing! Cause: (IOException) Error opening file './../giraffe_storage/universe/-111727825_817108$

Is this planet related to the thread's issue? Unfortunately the logfiles don't cover the same time and Nerva's file has no coordinates in it.

 

As said in previous post, this even occurs in a freshly generated universe (which I guess is then never shutdown before) So while knowing the server does not close down "cleanly" because of a restart script I made (because otherwise I get trouble with RAM usage and blablabla) I do not think that this is the problemcauser

 

mig

That's the world, nimmerland. I recognize the coords in the world db filename.

We tried removing that specific world from the universe files and allowing it to be re-generated, but it was re-generated with the same crash-causing problem.

If you want two fresh logs, covering the same time frame, one from each of us, we can probably arrange that.

EDIT: Keep in mind, Sme4gle and I are in different time zones. He's six hours ahead of me.

 

You can try the sigint command mentioned in the linked thread or there might be a real server command to exit the server cleanly in your script to prevent file corruption in general.

So the file opening error above was due to the deleted file and is not common to the error in general?
Yes to the fresh logs of the same timespan.

Edit: @Nerva no need to hurry.

 

No, the file opening error was due to the world file being bad, after it was generated. It had nothing to do with our deletion of it - the same error occurs on another planet in the same system, which we have not touched.

We removed the world and allowed it to be generated fresh to ensure this wasn't some fluke or corruption, or due to players placing modded or otherwise-bad items on the planet's surface. We did it to make sure that the world was being generated in a bad state, consistently.

Give me a bit to confer with Sme4gle, and we'll get you those concurrent logs.[DOUBLEPOST=1424557127][/DOUBLEPOST]Okay, got you the log.

 

Heres my server log like Nerva requested:

 

Just to make sure that it's not the dirty shutdown that's the problem, we're going to try a clean shutdown, and generate a fresh iteration of the world. Expect another pair of concurrent logs shortly.

 

New log after clean shutdown AND planet deletion, so basicly it has no chance to be a corrupt file in my opinion?

 

Here's my client log from the same crash.

 

The server seems to be acting normal, no IO Exception this time. At the time of Nerva's error, the server logs list a disconnect, probably because of the client has crashed. Then the world is unloaded as usual due to inactivity. Are you able to try with x64 client?

 

I'm not on a 64-bit OS. It won't even start.

 

This is hopeless, it doesn't work with whatever client, doesnt matter if they are 32 bit or 64, we tried everything now:

* "clean shutdown + planet removal"
* clean generated universes
* clean characters
* 64/32 bit players

This seems to be a problem with the world generation to me, and you are saying everything is working correct, but Nerva's client, well then explain me why all clients have this problem on countless planets.

 

Just the logfile of the server is not showing anything unusual. Also I did not blame Nerva's client / installation but the client in general. It still can be the server's fault (Either the server sends faulty things or the client can't stand the legit data from the server for those planets). There was a bug report on this kind of thing for version 665 and it might be useful to reply to this specific report with your own bug report on version 668. As there are many clients and many occasions where this happens I would suggest a rating of 3 for this bug. You may link to this thread or the ones below in your report.

I tried the search today... This looks like
http://community.playstarbound.com/index.php?threads/multiplayer-ctd-on-certain-planets.91749
and http://community.playstarbound.com/index.php?threads/random-planets-crash-game.92376/

 

my buddy and i have encountered the same problem on servers too, and it's not the entire planet, it's some regions that you cant go too close to it, otherwise the window will simply disappear and you get no error message or whatsoever, the game is simply gone.

 

exactly Joe, it sometimes seems to happen upon entering certain biomes aswell, its not allways entire planets.

 

Hey, glad I'm not the only one experiencing this issue.

I compared my log to Nerva's original log, and I get the exact same error message, verbatim:

Code:

[12:39:49.479] Error: Exception caught in client main-loop: (NetStatesException) Extra data at end of NetStepStates::load!
  StarException::StarException(char*)
  NetStepStates::readFullFrom(DataStream, unsigned int, double)
  NetStepStates::readFull(ByteArray, double)
  Monster::Monster(MonsterVariant, ByteArray)
  MonsterDatabase::netLoadMonster(ByteArray)
  EntityFactory::netLoadEntity(EntityType, ByteArray)
  WorldClient::handleIncomingPacket(shared_ptr<Packet>)
  UniverseClient::handlePacket(shared_ptr<Packet>)
  UniverseClient::update()
  ClientApplication::updateRunning()
  ClientApplication::update()
  StarApplicationBase::run()
  _SDL_main
  _console_main

Found a planet on a server I host that consistently crashes when moving near the surface. The planet is a cold planet. I start on a tiny island with a large ocean beneath it. Now, if I quickly jump into the water and sink to the bottom, my game won't crash. It only appears to happen when I stay up top for too long.

EDIT:

Per suggestion, here is the planet name and coordinates:

Name: Delta Zeta Hya 9527 I
X Coord: 229975993
Y Coord: -504427525