Double-adding a block
helkarakse opened this issue ยท 16 comments
Hi,
Just had a server crash twice, same crash log, same position. Pastebin as follows: http://pastebin.com/ih7fFzHa
Crash log says to post this in the issue tracker, so here it is. Reactor is a 5x5x5 setup, single fuel rod, three high internal dimensions, water surrounding the fuel core. Reactor is connected to a mekanism ultimate energy cube via reactor power tap.
Interesting. I've never seen this one before, thanks for the report. Was anything notable happening on the server before this happened? Looks like someone's walking into the area where the reactor lives, or logging into it.
To work around this, I suggest chunkloading that spot.
The area is currently chunk loaded with immibis dimensional chunk loaders (Server is running resonant rise 2.5.1). I noticed that when I am the first one to log in (after a crash) the reactor always crashes the server (with the same error).
Also, would it be possible for you to try loading your map up on a non-Bukkit/MCPC+ server to see if it also crashes?
I'm not the server admin. I'll refer the server admin to this issue for this matter.
Updating to version 2.5.2 (latest) of Resonant Rise has appeared to fix the issue. I think we can close it for now. Thanks for the help.
This is likely an issue with Big Reactors and threaded chunk loading in MCPC+, which would explain the randomness and why it doesn't blow up without MCPC+.
In the short term, can you chunkload your reactors? I'm working on a refactor that will make BR completely thread-safe, but it's a substantial rewrite.
I'm currently updating our RR servers to 184 which may be of some help for you regarding thread safe methods.
Disabling async chunk loading in 184 will likely fix this until I can get the threadsafe version done.
Crash has occurred in a different area under RR 2.5.2. http://pastebin.com/mSvZmF4a
Hi erogenousbeef, here is a world download with both mentioned region files. Load it up in SSP and TP to the coords to take a look. Unfortunately I can't provide the entire world as it is extremely large, but the region files should suffice.
https://dl.dropboxusercontent.com/u/82904851/BigReactors-issues-64/world.zip
Unfortunately I have no idea how to recreate the crash (it seems to occur randomly) and running the servers without MCPC for a long period of time would be detrimental to the server.
I've just put out 0.2.10X1 with a major multiblock rewrite. I don't suppose I could ask you to try it on a backed-up copy of your server with MCPC+'s async chunk loading enabled?
Haven't seen any issues yet but #71 has become a large issue.
I haven't seen this crash occur since updating to 0.2.10X1 (We are now on 0.2.10X2)
You can safely assume this has been resolved.
Thanks for the fixes!