Bug [SERVER][RAIL] collision time limit reached.

[RabidBat]

[SERVER][RAIL] collision time limit reached. Returning to last possible

Anyone know how to identify the location of the offending rail/entity causing the message.

Noticed quite often this message spams the server console. It's usually a docked entity that has bugged out causing the issue.

Only times I have been able to work out who or what is causing it is when there have been few players on the server and going through a process of elimination.

Cheers.

 

[therimmer96]

>implying messages in the console are worth anything

Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong
Waiting for Pong



That message has existed since rails have been implemented, you can ignore it, everyone else does. Looking for lag from the F7 menu is much more reliable than the console. Shit that gets printed to logs and console is more for schema's benefit than ours.

 

[Asypha]

I have come across only two possible causes for the "[SERVER][RAIL] collision time limit reached. Returning to last possible" message.

1. Two entities on the same rail are colliding with each other and thus being bounced back and forth. The fix is to set the rail 90 degrees from the normal travel direction when you want something stopped, and not to just rely on the collision detection to keep things in place. Also, the easiest way to check if this is the cause is to go into build mode and inspect the rail under any entities you expect to be in static positions. If the rails they are sitting on are pointing at dead ends, you should move onto the second option.
2. An entity was commanded to move along (or around) a rail, but was stopped by blocks in the way. If it is being spammed, then it is likely that the entity was trapped within another block at some point and is now confused about the correct way to exit the block. (For example, I had a door that was stuck half way into a static block, but because it didn't know which way if should exit, it just bounced endlessly.)

In both cases, an inspection of all docked entities will normally turn up the culprit. It will take a bit, but it shouldn't be too insane.

 

[AndyP]

Correct, thats the two causes for that message.
Sure the message itself could contain more information, and its impossible to track this down on larger servers.
However, as rimmer already said, there is a lot of log entries that only serve debug purposes and for examination of constructs.

This specific one for example is in since we had issues that two wedges could not pass each other on rails, and they virtually occupied more than one block. This way we can see its a collision blocking movement, and not power, mass-enhancer or input related.

I assume there will be a major clean-up of those when leaving the alpha.

- Andy

 

[Asypha]

I assume there will be a major clean-up of those when leaving the alpha.

While I look forward to that cleanup, it would be nice if a more detailed version of this message made it though the purge. It has proven helpful to at least know that there is a rail issue. We are currently only working in 5 sectors, but if a future message could call out the entity the rails belong to, it would be helpful even once we venture out.