Filler End Of Power World Corruption
Anarcraft opened this issue ยท 15 comments
I reproduced this with BC 6.1.5n and 6.1.8. I'm using the FPSPlus mod. (didn't test without it)
I place a TE hardened energy cell adjacent to the filler and when the cell runs out of power and the filler stops the world stops responding. My client doesn't crash I just can't interact with anything in the world or any extra inventories (like backpacks). There is no crash log. I haven't tested this with another power supply.
To fix my world I use MCEdit to delete the filler. Also I've never seen a corrupt block cause MCEdit to become unresponsive for a long time as it tries to render the info as you select it until now.
Hello there,
encountered the same problem with a filler connected directly to a culinary generator
from ExtraUtilities.
Game doesn't crash but I'm unable to open a chest, machine inventory etc.
Used modpack Direwolf20 V.1.7.10 rev. 1.0.2,
deactivated mods : Hat, Hatstand
extra mods : newer files for aromic backup (core and backup),
Aromic's Mining Dimension. BetterChests & Enchanting plus
and FastCraft V1.9
World still loads but shutting down won't work. Have to quit the FTB-Launcher via the 'X'.
Selecting the filler in MCEdit isn't possible. Try to delete the culinary generator
Tried the new version 6.1.8 of buildcraft but still got the issue.
Hope this description helps to solve the problem.
Thorton101
Edit#1: I ran a small self-created modpack at the end of august this year
with a buildcraft version 6.0.18. I used there the filler and i worked
correctly. Also the 'Hermitcraft Modsauce'-Mod on AT-Launcher has
a buildcraft V6.0.18 in it and there the filler worked with no such problems.
I wasn't able to reproduce the tick/lag problem but when I depleted a Leadstone Energy Cell using a filler, I couldn't get it to start charging again. Energy was traveling to it but it wouldn't charge even though the filler wasn't working.
The only way I could get it to work again was to destroy it and place it again. After that I couldn't reproduce it until I restarted the world (it was a single player world by the way).
Another thing I saw, was that even after the filler completed it's work (the texture was turned from green to red) the energy cell continued to output energy into the filler. The filler then had enough energy to clear 539 blocks, perhaps it's internal energy store should be reduced.
Also a bug in Thermal expansion, I couldn't disable the energy cells using redstone signal.
@psxlover - fillers taking energy when not working is a perfectly valid bug. Also, the internal buffer should indeed be depleted.
Does it happen in the overworld or does it happen in Mystcraft/custom dimensions?
I'm using BC 6.1.8 and I'm having maybe the same issue. On one of my fillers, the TE energy cell went empty. After that I have tried to break the energy cell, the two EnderIO conduits connecting the energy cell and the filler and the filler itself. Breaking the energy cell and the conduits took a long time, but after they were broken I could play without problems. But breaking the filler isn't possible. The server just looks up until I log out. When I reenter the world and supply the filler again with power I can break it normally. According to the java profiling tools, the lockup happens in one of the NBT methods. I forgot to take a screenshot, so I can't say which method exactly it was. And I wasn't able to reproduce the issue since I broke the faulty filler.
The issue seems to only happen on SMP, on SSP I have never encountered that problem.
I hope that helps a bit to find the problem.
Edit: Just tested it a bit more on SSP and could get it to also happen there.
@Dav92 and others - if any of you reproduce this problem, get a java profiling dump! This is likely the issue that affects fillers, builders and quarries.
Ok, I've succesfully reproduced the issue and profiled it with jvisualvm. Hope that helps a bit:
http://www.mediafire.com/download/636bzy469grcvy0/profiled1.apps
Is the following the procedure while profiling:
-Start profiling
-reproduce the issue
-take a snapshot and export it
Or is there anything else to do?
This happened in the overworld on SSP. I tried a quarry and did NOT have the issue. If you also need me to do a java dump I can, just never done that before so I would have to figure it out...
@Dav92 - This is for profiling, but what you should do is also make a few Thread Dumps.
Ok, I've reprofiled it and took a couple of snapshots. I also made a couple of thread-dumps. Hope it helps:
http://www.mediafire.com/download/va6m916azek09p3/mc-bc.apps
What I also found out is, that you can just wait and the problem fixes itself. In my tests it depended on the size and progression on how long it takes (a 60x60x50 filler with ~10% done needed around 10minutes, I haven't had the patience to wait for a filler with the same size but ~95% done).
Is there anything else that can be done to help? E.g. trying to reproduce it with as less mods as possible?
Edit: Like in the logs from @benten2000 in the thread dumps you can see that there is waila needing very much processing time.
Had the same problem today using the filler then getting long loading times to access any UI (can't keep up errors from 24'000ms to 77'000ms).
Was using the FTB Direwolf20 1.7.10 pack v1.0.1 and tested v1.0.2 (SSP) and found it the same.
After some testing with a few mods at a time I found that Waila (at least so far) was causing the lag.
The default installed v1.5.4a and the latest v1.5.5 would cause the same effect.
The error log at least this time showed something (if any help).
I just tested the fillers with the minimal amount of mods (just BC, TE for energy storage and EnderIO for conduits) and there the problem wasn't occuring. So it could be just the "communication" of Waila with the filler or something along these lines.
I also have this problem. SSP 1.7.10 with DW20 FTP pack as base, have removed some mods and added a few and even updated a few. Also tried it on on FTB Light same problem.
I managed to break the filler (problem "gone") and most of my world is ok as far as I can see.
No useful stuff in the logs I'v checked so far.
I did a small flat "horizon" fill and that managed to finish with huge lag. When I then tried a bigger fill it happened.
Oh, yes I am using 6.1.7 so far, will now try 6.1.8
I was thinking about this bug more because now it happened on our server. If it's a WAILA bug then is WAILA modifying the world block? I ask because this world block (Filler) also caused MCEdit to become unresponsive when looking at it (MCEdit has a similar HUD with info of what is being looked at). Or do both WAILA and MCEdit suffer from a similar bug?