Solderer dying
raoulvdberge opened this issue ยท 45 comments
I should note a Twitch viewer commented that he had to break and replace various other units and cables after leaving and coming back, so it's probably not restricted to just the solderer. Did he bother filing a bug report for it rather than just complaining about it in a Twitch channel? Don't be silly. XD
@RealGrep Can you try this debug build in a fresh world with just a solderer? Let me know what the console says.
refinedstorage-debug.zip
Can't seem to get it to freeze up in a fresh world with only the solderer, at the moment. Doing the same thing as usual and going to the nether, waiting a couple of minutes, returning and hitting the redstone control mode button. Do you still want the logs? Shall I try it in my normal world where it locks up?
Spoke too soon. Managed. I wonder if it has anything to do with TheOneProbe. It's spewing errors there. If not, at least the NullPointerException might be a clue. Anyway, I zipped up the entire log. To find the relevant section of the log, search for a few asterisks, i.e. "***".
https://www.dropbox.com/s/ltpe188zevnlj51/fml-server-latest-rs1.zip?dl=0
So basically, I moved it away from the spawn chunk to make sure it wasn't loaded when I left. I put the nether portal slightly away from the solderer (two chunks in one direction, one chunk over in the other axis). Went to the nether, waited a couple of minutes, and returned. The redstone mode button no longer worked. Solderer seems to stop ticking.
There's a possibility the TOP NPE is from this: McJtyMods/TheOneProbe#14 . And if so, that might fix it. Though the solderer isn't really multi-part, so not sure this applies. I think it's trying to show the contents of the solderer and failing.
EDIT: Probably not. Error seems to be as it's trying to get the contents using capabilities. Perhaps it's not implemented correctly in RS?
Huh:
[14:22:01] [Server thread/INFO] [STDOUT/]: [refinedstorage.tile.TileMachine:func_73660_a:54]: [REFINED STORAGE DEBUG] Machine Tick 1280 (Connected=false,controller=null,pos=BlockPos{x=517, y=64, z=-210})
[14:22:02] [Server thread/DEBUG] [FML/]: The world 1a53be82 (Addle's Realm) may have leaked: seen 140 times.
**************** 14:22 go to nether, solderer so far ok
Your solderer wasn't even connected when you went to the nether?
lol, that is true... I actually need the solderer to be connected to a controller (which is not technically a machine) before you go the the nether :)
This should fix The One Probe error:
This is the log with the solderer connected to a creative controller:
https://www.dropbox.com/s/addzyr6l6wmqeq9/fml-server-latest-rs2.zip?dl=0
What happened here?
[13:58:23] [Server thread/INFO] [STDOUT/]: [refinedstorage.tile.TileMachine:func_73660_a:54]: [REFINED STORAGE DEBUG] Machine Tick 1500 (Connected=true,controller=refinedstorage.tile.controller.TileController@7a5f83ce,pos=BlockPos{x=126, y=64, z=293})
[13:58:24] [Server thread/INFO] [STDOUT/]: [refinedstorage.tile.TileMachine:func_73660_a:66]: [REFINED STORAGE DEBUG] BlockPos{x=126, y=64, z=293} My state changed to false
[13:58:24] [Server thread/INFO] [STDOUT/]: [refinedstorage.tile.TileMachine:func_73660_a:54]: [REFINED STORAGE DEBUG] Machine Tick 1520 (Connected=false,controller=null,pos=BlockPos{x=126, y=64, z=293})
Did you change the redstone control of the controller?
Also, check if the error still persists with the other debug version I posted.
Yep, before I go to the nether, I click it to make sure it's ok. I click it again when I return to see if it's frozen up or not.
About to test with the new version.
Ok, same basic result, except the errors from TOP are gone (so as I figured was likely, they are unrelated, but still nice to have them gone):
https://www.dropbox.com/s/1c7xupl88pw9n4a/fml-server-latest-rs3.zip?dl=0
Is it me, or did it actually change the redstone control mode when I click the button? The button's displayed state certainly didn't change.
Still in the game next to it, and I see no ticking messages. The last line on the console is still: "[16:18:11] [Server thread/INFO] [FML]: Unloading dimension -1"
Interestingly, the redstone control button seems to work ok on the creative controller. No messages are generated in the console when I click on it, but it updates and changes modes.
Does the redstone button work on other machines like the grid or importer after coming back from the nether?
Nope, I put down a ton of machines. None of the buttons work in any of them, except the creative controller.
https://www.dropbox.com/s/xz2txil4olqe4xm/fml-server-latest-rs4.zip?dl=0
What I find interesting is it seems things are ok until the Nether is actually unloaded, at which point all the messages from RS stop.
Yep, the two are related somehow. What I did was put down a chunkloader in the nether to keep it loaded (The ExtraUtils2 one, along with a bunch of lava generators to power it).
Did the same test as usual and waited a couple of minutes in the nether. Came back, and all the machines seem to be responsive, and I can change the state of all the buttons. The machine tick messages continue, and no unloading dimension -1 message appears.
That's rather weird!
So if you keep the nether loaded while the machines are in the overworld, everything keeps working? O_O
Odd, I just repeated it, the nether didn't unload, but the buttons are unresponsive. Might be a red herring.
Just flew about 300 meters out and came back, the machines are working again. In case that helps.
Ok, I removed the check if redstone is enabled for testing. Can you retest with this?
refinedstorage-debug.zip
Same basic test sequence with new debug build: https://www.dropbox.com/s/ypvusruh0p2d1rd/fml-server-latest-rs5.zip?dl=0
Here you go: https://www.dropbox.com/s/wh34e26rlh1j64e/fml-server-latest-rs6.zip?dl=0
And you're welcome, I'm glad to help out. :)
Okay, can you now try with this build:
Sadly, bug still exists. First test worked ok, so I stayed a bit longer in the nether before returning. This time, the redstone mode buttons are unresponsive again. Sorry. :(
Really weird issue. Will definitely look at it 100% at the weekend, considering as this is the only remaining bug open.
It is. Seems to be better if I keep my area chunk loaded, so definitely seems to be related to chunk loading/unloading.
And yeah, pretty much the only real bug I know of left in it. heh
Oh, and I can definitely package the mod pack and test world I have for you and put it somewhere you can grab it. Just need a private channel for me to give you the link. Maybe I can temporarily add my e-mail address to my Github page for you or something.
I am in my world seeing a more generalized version of this problem. Sometimes if I am off roaming the world or in another dimension I come back to my base and the controller does not see any of it's connected devices. At first I thought it was a power issue, but after putting a huge capacitor to make sure it wasn't fluctuation in my power grid... nope.
It's easily fixed by going to the controller and toggling the redstone mode so it goes off then on, where it rediscovers all it's devices. However this is bad news for unattended automation.
All chunks involved are chunk loaded with Ender Anchors (random things) in an attempt to resolve this. Didn't help.
Possibly (probably?) related: the controller is in a different chunk than all it's devices other than it's power source. All should be chunk loaded.
I know what the cause is of this issue. I could do a simple hack to fix it, but I'm reworking the machine system once more, so it'll be fixed as soon as the rework is done.
That's good news! Congrats on finding it. :)
Not a problem, best to fix it properly, IMO. In the meanwhile, keeping it chunk loaded seems to help.
Should be fixed now! @RealGrep, please hang in there. Will post a test build for you when the dev branch is a bit more stable.
Well, did the usual test sequence three times, and I can't replicate the bug any more. So looks like it's fixed, unless I was just unlucky in my testing. Good job! :D
If you DO see it again though, please do not hesitate to make me unhappy again xD