Kerbal Space Program

UPDATE: I added and deleted an alarm from Kerbal Alarm Clock, and the problem disappeared. However, when I moved another ship closer to this space station, the warp problems reappeared. I tried ramming the new spacecraft part into the station, which did nothing, although turning the ship caused the entire station to promptly explode into fragments. Then, the parts still connected to the command module I was using had their rotation locked and started accelerating on their own away from Kerbin. The acceleration always seems to occur AWAY from the orbiting body.

I am experiencing the same problem.
I'm not sure when I first encountered this, probably it was early in 0.23.5. That was just after I started using mods and after removing them (especially the "Distant Objects" mod seemed to be related) the problem was gone (that's why I didn't report it back then). But now, in 0.24 it is back, with or without the mods.

To me the ships' behavior looks like they are not following their orbit anymore but just moving along their current direction vector. The orbit is following the ship, hence being shifted along the travelling direction, keeping its relative location with regard to the ship.

Further, strange physics forces are applied to the ships sometimes, causing them to be twisted around in an unhealthy way and (sometimes) ripping them apart.

When you return to the Space Center (losing data since last save) and then jump back into a ship, the problem is gone until you start time-accelerating again. Physics warp works just fine and doesn't trigger that problem.

Quitting and restarting the game seems to resolve that problem at least for some time but eventually it occurs again. I haven't found a way to reliable reproduce this issue so far but I'll keep investigating.

I uninstalled Kerbal Alarm Clock and the bug persisted, so it wasn't that mod at all. My issues were also definitely space-station related, as I haven't yet reproduced this without visiting a space station in-game, and the bug didn't appear until shortly after I discovered how to use docking ports.

@terbelain 's method reliably reproduces the bug for me. I think it has something to do with the fact that the ship was undocked from the space station. Mine is made up of one asteroid and a handful of parts attached during different launches. It tends to occur immediately after undocking from said station. I haven't had it occur after undocking two simple ships.

Another odd thing to note: SOMETIMES the space station accelerates to infinity with me, particularly when it's closer, and sometimes it does not.

This also happens to ships that were never even close to my station.
My current guess is that the physics tick might get out of sync with the game tick due to bad performance while being around the station (or large structures in general) and tries to catch up on during time-acceleration where physics are usually disabled. Or it might be something completely different. ;)

I just noticed that this is probably the same glitch:

https://www.youtube.com/watch?v=ul6FXQMT1GI#t=192

well, the effects of it at least.

Happens on Mac/OSX as well

Still happens in 0.25. Just checked it in a stock install with the savegame and reproduction steps from comment #5.

I just want to add that I am experiencing this exact behavior (without asteroid) in 0.25 on linux x64, however the sfs it is happening in is using some mods. If it's useful to load a save that was created with mods into a stock instance, I can definitely do that. I can probably edit out all of the non-stock parts. Either way, the behavior is the same, and I can trigger it by loading a particular save and using time warp. The root node of the active ship will have its orientation locked and start accelerating away from its SOI body.

Edit: I did prune the non-stock parts and have replicated the behavior in stock 0.25. I'll attach two sfs files, approximately 180 s apart, one before the bug shows up and one after. A space plane was drifting towards a refueling station, and I was holding down 'n'. The problem seems to be related to the two vessels being close together.

Squelch wrote:

I've marked this as needing more information for the simple reason that the reproduction steps require a complex savegame. (The possibility of a corrupt save cannot easily be ruled out)

Could the before and after savegames in http://bugs.kerbalspaceprogram.com/issues/2753#note-14 be used to rule out or isolate save corruption?

lomendil wrote:

Could the before and after savegames in http://bugs.kerbalspaceprogram.com/issues/2753#note-14 be used to rule out or isolate save corruption?

The short answer is not really.

The long answer is that by your own admission, it is a modded game, and that you have manually edited it. While this may work in many cases without problem, we are dealing with an esoteric bug that must be concisely reproducible to be able to forward to the devs. Is the problem down to some whitespace or improper line ending in the save? Was that whitespace or line ending introduced by your edit or the previous mods? just for example.

I will review your saves in the hope of discovering a step by step reproduction. Thank you for the time spent on this.

We must rule out as many variables as possible, including saves if at all possible.

Could you please explain what kind of reproduction would be appropriate? As the problem seems to be related to complex crafts it is probably unfeasible to not include a savegame in that process.

Just for the record (because I didn't mention that earlier): the savegame from #2753-5 was directly taken from the game and not modified manually. If it is corrupt this must be due to another bug. ;)

tirbelain wrote:

Could you please explain what kind of reproduction would be appropriate? As the problem seems to be related to complex crafts it is probably unfeasible to not include a savegame in that process.

We need to remove as many variables or unimportant details as possible. This bug is not occurring for everyone all of the time, so there must be a unique set of circumstances or steps that will enable us to reconstruct it. If it does indeed require more complex craft or combinations of parts, then in all probability those combinations will be a factor. To forensically examine complicated craft and scenarios for no reason is a waste of time, and is of no use to the developers in identifying the cause.

The steps need to be as simple as possible. ie

Put vessel a into x orbit.
Put vessel b into Y orbit
Switch between vessels A & B once
Increase Vessel A orbit to Z

The vessels must have the minimum number of parts to produce the issue.

Just for the record (because I didn't mention that earlier): the savegame from #2753-5 was directly taken from the game and not modified manually. If it is corrupt this must be due to another bug. ;)

Indeed, and this is a possibility that can't be ruled out and would need investigation. There is no doubt in my mind that this is an issue, but the reasons for it are very unclear, and there are too many other factors that could have caused it. We need to rule them out to be able to justifiably call it a bug. Thank you for your efforts on this problem. I'm sure we'll get to the bottom of it.

The edited savegames are simply unusable. Upon loading they turn KSP into a slideshow, and the logs are so full of errors, that nothing meaningful can be obtained.

I have removed them from this issue so they won't confuse others. Please refrain from departing from the excellent bug reporting guide found here

After spending the day (more or less with no avail) on trying to find a reproduction from scratch, at least I think I found a way to eliminate the savegames (kind of) and every single spaceship part from the equation. Here's what I did:
- reproduce the glitch just as before by using the steps from #2753-5 -> when the problem occurred once, the game seems to be broken
- go to the main menu and start a new (sandbox) game
- build a tiny rocket and launch it into orbit
- again activate 10x time acceleration and see the effect happening on your tiny craft in a fresh savegame

The effect can be observed even on a one-part object. I tried it with a probe body and a battery (in two separate launches) that I put on top of a decoupler, decoupled and deorbited the rocket (to get rid of any other objects in orbit) and jumped back to the part (via tracking station).
Actually you can repeat this as often as you like once the game is in that broken state, either with the same game or with a new one. It will remain there until you restart KSP.

I know that this is not ideal because you still need the savegame to break it. But it shows that the problem comes from a global game mechanic (like the orbital physics system that is probably initialized only once and keeps its state across savegames). Unfortunately it does not show what causes this problem in the first place.

Actually I managed to break a new game without using saves but that took me several hours of uninterrupted game play. Restarting KSP (hence resetting the problem) and trying to reproduce it again with the last steps from the previous session didn't work.

I am also experiencing this exact issue. It seems like the vector that the craft is moving on just stops updating so it just continues moving at it's exact speed and direction without gravity affecting it at all. It's almost as if All gravity in the kerbal system instantaneously disappears so you just but only in relation to your current space craft. However I think this may have something to do with the SAS system as it appears that only certain parts can bend after the bug takes hold. usually everything below the SAS module can bend if you exert force using RCS, but the SAS and everything above is rock solid stuck facing the EXACT direction it was last in.

I have also noticed something a bit weird when the issue starts happening. I was working on capturing some asteroids to start making a base and so I wanted a more stable capturing platform, so for the nose of my new asteroid capture craft I flipped a quad adapter upside down so I could put 4 grabbers as the nose. (It was an experiment that worked decently) However when the bug started showing up the grabbers would start shaking and vibrating and sliding around on the quad adapter. They would get all mashed together and not even be sitting on their connection points. Sometimes if I turn off SAS, RCS, and stop all acceleration of any kind they will stop moving around and just sit their in their new messed up positions.

tirbelain wrote:

The effect can be observed even on a one-part object. I tried it with a probe body and a battery (in two separate launches) that I put on top of a decoupler, decoupled and deorbited the rocket (to get rid of any other objects in orbit) and jumped back to the part (via tracking station).
Actually you can repeat this as often as you like once the game is in that broken state, either with the same game or with a new one. It will remain there until you restart KSP.

I know that this is not ideal because you still need the savegame to break it. But it shows that the problem comes from a global game mechanic (like the orbital physics system that is probably initialized only once and keeps its state across savegames). Unfortunately it does not show what causes this problem in the first place.

Actually I managed to break a new game without using saves but that took me several hours of uninterrupted game play. Restarting KSP (hence resetting the problem) and trying to reproduce it again with the last steps from the previous session didn't work.

Thank you for your observations. The problem persisting between saves is of particular interest, and may link in with some other reports that are hard to pin down, and go away after a game restart.

x5060 wrote:

I think this may have something to do with the SAS system as it appears that only certain parts can bend after the bug takes hold. usually everything below the SAS module can bend if you exert force using RCS, but the SAS and everything above is rock solid stuck facing the EXACT direction it was last in.

The SAS system and how it operates has been revised in 0.90, so may have a bearing on this problem

I have also noticed something a bit weird when the issue starts happening. I was working on capturing some asteroids to start making a base and so I wanted a more stable capturing platform, so for the nose of my new asteroid capture craft I flipped a quad adapter upside down so I could put 4 grabbers as the nose. (It was an experiment that worked decently) However when the bug started showing up the grabbers would start shaking and vibrating and sliding around on the quad adapter. They would get all mashed together and not even be sitting on their connection points. Sometimes if I turn off SAS, RCS, and stop all acceleration of any kind they will stop moving around and just sit their in their new messed up positions.

The Claw does seem to figure in several reports, but when tested in isolation seems to behave in all but the most complex situations. It has been suspected as the trigger, but similar behaviour has also been noted where no Claw is involved. It's a prime suspect, but only with circumstantial evidence.

I am also experiencing this on a regular basis, I'm going to try to narrow down the cause be elimination in my vanilla install, but it's a tedious process. Some symptoms I've seen:
- this can be the only ship active in the savegame, no rendezvous is necessary, or even any other flight anywhere.
- main symptom is ship going into a state where it can't be controlled, claims it's under acceleration, and does not obey physics. this usually happens coming out of a time warp.
- on stack separation, remaining upper stages go into wild rotation, but command pod seems to maintain its orientation, quickly breaking off the stack and going into the above state.
- ship often explodes on stack separation, probably just collisions due to above spin
- when ship is in stuck state, if you let it fly for a bit, it eventually gets stuck at a point in space entirely. for example, one got stuck in a state that its false-accelerated path was still suborbital, but it froze at 99998 meters and just stuck in space at that altitude.
- (this may be mod related, haven't tried it on my vanilla install, but coincides with this bug's other symptoms often on that savegame, and happens on stock parts) symmetrical decouplers only decouple one of them when all are activated in staging or action group

Running on linux x64

Certain conditions with the universal storage mod will cause this kraken. I've notified the maintainer to investigate. However, using that might be a good way to easily reproduce and see what the game engine is doing wrong when it occurs.

I am experiencing the exact same bug. I grabbed an asteroid, put it into orbit around Kerbin and now all ships that achieve orbit around Kerbin starts accelerating during timewarps. I am playing KSP through Steam without any mods and I have the newest version (1.0.4) installed.

@RexRectumIV

If you would care to share your savegame and logs, it would help us determine the cause? This is possibly the first report of the problem since the persistence bug was addressed.

nikvolf wrote:

But I have savefile where it can be reproduced regardless
MODS: Engineer, Protractor

Thanks for reporting, but unfortunately we cannot work with modded games due to any other variables they may introduce, and simply trying to keep up to date with them all. Stock reproductions only please.