This repository has been archived by the owner on Feb 27, 2024. It is now read-only.
Async Light causes IndexOutOfBoundsException SPacketChunkData #2405
Comments
limbo-app
added
status: needs triage
|
Is this related? #2276 (comment) |
|
Possibly but unsure. Another note to add is that I had to sigterm the java process. It just let the server hang but marked "connected" players as afk. Don't know if that's related. |
|
@sfxworks You're using an extremely outdated build of Sponge - you should update to the latest version. |
|
Upgrading the version revealed the chunk but it still occurs ]: SpongeForge |
This comment has been minimized.
This comment has been minimized.
|
Anything on this, @sfxworks? |
|
The recent update had this still occur. It gave me chunk coords. The ones specified revealed no active blocks other than ones rendered. Have not had a crash lately though. |
|
This is happening about twice a day for me. All my locations are pretty much random too. |
|
https://gist.github.com/sfxworks/00a6d57c373a26c873b0763036a83180 Had a new one today. Redacted coords. Staff to investigate affiliated chunks. |
|
Good news. We have continuous replication results here. We have a few region files that are giving similar coords and a player joining a planet in ExtraPlanets that has continually crashes the server. I can send anyone the world if necessary. |
|
Yes, it's happening for me daily. I had a client experience roughly the samething where a packet was too big and would disconnect them. Come to find out it was massive amount of EnderIO conduits sending tons of data to the client. As, I do not have ExtraPlanets, do you have by chance EnderIO? Thanks, hopefully we can narrow this down and help find a solution for Aaron. |
|
https://pastebin.com/raw/xfEudesN Happening 1-2 times a day now. |
|
Since you have the chunk coordinates, what region is that chunk at? It would be probably indicative of possible bad chunk data overflowing the packet. |
|
Thank you for looking into this, gabi. The chunks/regions where this happen seem to be completely random. Even happening in flat worlds with no active players. Is there anything debugging/inspection that can capture whats maybe sending the bad chunk data? Should this be opened with Forge, as-well? Thanks. |
|
Can either test and give new logs for any changes from this jar? spongeforge-1.12.2-2705-7.1.0-PACKET_BUFFER_TEST-1.jar.zip I know for sure it doesn't mitigate the issue but it does provide more information about how and maybe when something is going wrong. There's more validation I can write for previous writes to the buffer, but I don't feel that it is necessary. |
|
https://pastebin.com/raw/0U3tV4qS That's with build 3440. |
limbo-app
removed
the
status: needs triage
|
Sponge build 3442. |
|
This just happened to only me by teleportation to a already generated chunk. This has nothing to do with on chunk generation only happening as I thought and has nothing to do with CoFH Thermal Foundation. |
|
gabizou
added a commit
that referenced
this issue
Sep 28, 2018
Signed-off-by: Gabriel Harris-Rouquette <gabizou@me.com>
|
Ok, a new build of SF should have the chunk packet debugging info that would probably show us more of the issue, build number 3468 should be the one to do it. Please provide new logs if and when the crash/exceptions occur. |
|
I'm having the same problem on my server as well and it's causing a lot of trouble. |
As I just commented, a new build with some better information is Edit: SF 3468 is available on spongeforge downloads from sponge. |
|
@sfxworks @Besharia @BrandHyde @OldKingisdie Can you all load up all your crash reports that you have had with this problem and send them We are trying to narrow it down to a coremod, I'm going through testing them right now, you can contact me on sponges discord |
|
Yes, spent 5-6 hours yesterday trying to reproduce it in a manner that would be helpful but coming up short. Although, I do believe you are on to it there. The world I had a massive increase recently with this crash did get new additions to generation and glowstone was added at high rates. If it is due to light calucations that does make sense why it's happening more. I will boot up a test world right now with way higher spawn rates on glowstone and see what happens. The only other world I had higher crash rates with was the Nether, lots of light calculations in the Nether. Will report back! |
|
crash-again |
|
Yeah, no faster reproduction with the world filled with much higher rates of glowstone. |
|
I'm also having this crash 2-3 times a day. |
|
Question: Can someone try with async lighting disabled? |
|
I can do this on prod for you, it's the only reliable reproduction I can do with Aroma1997 world + players running around in there. |
|
@Slayer5934 read the issue comments, it’s caused by enabling a sync lighting. No need to post more logs until someone goes and posts that more testing with potential fixes is needed. |
|
We disabled async lighting, seems to not be crashing so far. We triggered it before by teleporting to different planets (you cant launch rockets because of #2501) but now it seems its not happening. Ill comment back if it ever happens again, let me know if youd like any other testing done. |
phit
changed the title
|
Any news about this issue? |
|
Haven't been able to rewrite async lighting, or test some code to fix potential async lighting issues. I've went and reworked how worlds are keeping loaded however, so there are some things to test. If someone does run into the issue again, please comment with which builds etc. |
|
I got this error on my server too. I know this isn't a tested Forge version for the Sponge version but because the crash reports are the same I thought it best to let you know. |
|
Get this crash today. |
|
Also had the same crash This one occurred when a user jumped into the Nether |
|
I don't know if this help but recently i just got this crash on versions
|
|
Should be fixed with SpongePowered/Sponge@fd9dfe0 Please retest with the latest build. |
Cybermaxke
removed
the
status: investigation
|
Okay i had this fix on to my server for about 3 days now also had my async lighting enable and i haven't got a crash yet. I but bare in mind i do not know how to reproduce the issue. |
|
@runescapejon That's good news. The issue occurs in rare cases when chunks lighting is being updated and simultaneously sending packets to the client. So there's not a easy way to reproduce. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
I am currently running
Minecraft: 1.12.2
SpongeAPI: 7.1.0-SNAPSHOT-d453c3e
SpongeForge: 1.12.2-2705-7.1.0-BETA-3206
Minecraft Forge: 14.23.4.2705
openjdk version "1.8.0_181"
OpenJDK Runtime Environment (build 1.8.0_181-b13)
OpenJDK 64-Bit Server VM (build 25.181-b13, mixed mode)
CentOS Linux release 7.5.1804 (Core)
Plugins
Mods:
Issue Description
Last I heard a user was in space looking at schematics in chests and then they crashed... I am usually able to see what's up based on the crash report but the only evidence I have is that player statement and the dim affected in the crash report. Seems like it may correlate with extra-planets. Unsure.
https://gist.github.com/sfxworks/2a3d0f1a51d2fd5294b7b31203469a5c
The text was updated successfully, but these errors were encountered: